Automated detection of political ideology from text: a case study of newspapers in Uruguay
Although objective reporting is the landmark of professional journalism, several academics have argued that the media is ideologically biased. The association of the printed press (newspapers) with political parties have long been acknowledged in Uruguay. However, the lack of studies that empirically demonstrate and measure the extent of ideological bias towards political affiliations has prevented scholars from addressing the ideological diversity of newspapers. Here, the author describes the use of natural language processing and unsupervised machine learning algorithms in conjunction with network graph analysis to investigate the political leaning of five newspapers that published a total of 530 news articles on two political candidates from opposing parties during the election cycle of 2019.
THE INFLUENCE OF MEDIA IN SHAPING PUBLIC OPINION_
The media -radio, television, press, and online- has a vital role in society and its main function is the communication to the public of local, national and international events. Media could certainly influence what people think about a range of national and international issues. Several factors can contribute to the degree in which people are influenced by media; and these factors vary according to geographical and cultural contexts. For instance, a recent study co-authored by Hong Tien Vu and Peter Bobkowski from Kansas University showed that the strength of the effect of journalism on public opinion (from 16 countries on five continents) was dependent on the public's age, educational level, living area, political ideology; and also on the nation's macro variables such as economic development and media freedom [1, 2]. The study found that agendas set by media showed a moderately high correlation with issues the public considered most relevant, with countries such as South Korea, Taiwan, South Africa, Philippines, Mexico and Chile displaying statistically significant relationships between media and public agendas .
Media can also limit the scope of arguments and perspectives that inform public debate, and the subsequent construction of not only public belief but also their attitudes towards social change; as it was demonstrated by Happer and Philo in 2013 . These authors could show that for the issue of disability in the United Kingdom, an increase in reporting from printed media that discussed the topic in unsympathetic terms led to negative public opinion on disability benefits and the persons who have them. Interestingly as well, the authors showed that repeated exposure to media messages related to climate change pre-disposed the attitude and behavior of people in adjusting their views and opinions to new information . This means that public opinion could be selectively changed by exposure to media; a phenomenon that in United States was recently shown to occur when university researchers intentionally intervened 48 media outlets to activate public expression, causing citizens to discuss major issues of policy and politics as part of the ongoing collective 'national conversation' . Interestingly, the authors could detect that their media intervention altered the composition of Twitter opinions expressed in the national conversation by 2.3% towards the ideological direction implicit by their published articles; and an increase in public engagement and discussion on Twitter that was 62.7% higher relative to the media's day's volume .
Overall, these works described that in different countries media had a considerable and tangible effect in shaping public opinion on public policy in general and on political issues in particular. This raises the question if media effects on public opinion is intentional or not; and more importantly, if press and media may or not reflect reality but filter and shape it instead according to their biases (inherent or intentional) towards certain policies and political views.
IS MEDIA BIASED?_
Although the premise of professional journalism rest on objective reporting, several scholars have described that the media consistently displays ideological bias [5-8]. Ideology is described here as a system of ideas, beliefs and ideals which constitute the basis for political and economical theories that guide policy making . Media bias is considered intentional if it results from both a conscious act or choice and is sustained over time [5, 7]. In this sense, media bias is considered to be a systemic tendency instead of independently isolated incidents in which either journalists and/or media owners purposely implement in order to obtain a concrete political, social and/or economic goal.
Three types of media bias have been described , and these are: COVERAGE, GATEKEEPING, and STATEMENT. COVERAGE bias refers to the visibility of topics and entities, such as a person/politician or country, in media coverage. GATEKEEPING bias, also termed selection bias or agenda bias, refers to which stories media outlets select or discard for reporting. STATEMENT bias, also denominated presentation bias, relates to how articles and stories choose to report on concepts.
Media bias in news content could significantly impact the political attitude of voters and thus influence the outcome of elections . What then are the potential effects of biased news consumption? A likely outcome would be the reduction of political diversity and views across the population; which in turns would diminish freedom of expression and democratic values. Indeed, it has been suggested that presence and cultivation of ideologically diverse news content would lead to healthier democracies . Because of this, several countries have established laws that regulate media ownership as means to limit concentration of media outlets owned by few individuals or groups and their associated political ideologies [10, 11]. Concentration of media ownership has been the status quo in Uruguay [12, 13], a country located in South America that is the subject of study in this work.
IS THE MEDIA IN URUGUAY IDEOLOGICALLY BIASED?_
According to Adolfo Garcé, a political scientist, the media in Uruguay has been historically biased in political matters and closely associated to political parties until the 1950s, and gradually acquired more independence over political views ever since . Following a period of dictatorship in which many media outlets were censured and closed, Uruguay re-established its democratic government in 1985, to which the media and in particular the printed press adjusted its political preferences onward to match the new ideological climate of re-stablished political parties when democracy was re-gained in the country . Since 1985 then, the primary type of journalism practiced in Uruguay is 'declarative' in nature; and thus mainly characterized by reporter's citations of statements pronounced by politicians . Because of this, the ideological perspective -specially in the printed press- of a media outlet could be evident by the analysis of the nature and frequency of citations derived from statements given by political figures who are aligned with the ideology of the media outlet in question. The author himself has been noticing ideological bias over the years towards a particular politician when reading the political section from one of the major newspapers in Uruguay. This prompted the author to engage on a research-based art project for the automatic detection of political ideology from text using machine learning and natural language processing algorithms.
The objective of the current work is to implement automated techniques to identify political ideology from text, mainly machine learning and natural language processing algorithms such as topic modeling and clustering. For the purpose of this study, text constituted 440 newspaper articles written about two major political candidates running for president of Uruguay that were published by journalists from three different media outlets during the period between June 1st to July 25th of 2019. Text also included an additional 90 news articles written on two political debates held on October 1 and November 13 of 2019. The author also included documents containing the programatic outlines being championed by each candidate and their respective political parties.
BRIEF INTRODUCTION TO THE POLITICAL CONTEXT OF URUGUAY_
Uruguay implements a presidential form of government with division of power among the executive, legislative and judiciary branches. The president, who is also the head of the state, is directly elected by the people for a five-year term. It is the president who then appoints a council of ministers for each administrative department . The vice-president oversees the national legislature, which is composed of a bicameral parliament also elected by the people for a five-year term, and is composed of a 31-member senate and 99-member chamber of deputies. The last national elections were held on October (1st round) and on November (2nd round) of 2014 in which Dr. Tabarè Vàzquez -representing the leftist political party known as Frente Amplio- was elected president. The current presidential and parliamentary elections took place on October (1st round) and November (2nd round) of 2019. For this election cycle, there were eleven candidates running for president representing the following political parties:
Daniel Martinez > Frente Amplio (incumbent party)
Luis Lacalle Pou > Partido Nacional
Ernesto Talvi > Partido Colorado
Guido Manini Rios > Cabildo Abierto
Gonzalo Abella > Unidad Popular
Pablo Mieres > Partido Independiente
César Vega > Partido Ecologista Radical e Intransigente
Edgardo Novick > Partido de la Gente
Gustavo Salle > Partido Verde Animalista
Daniel Goldman > Partido Digital
Rafael Fernández > Partido de los Trabajadores
The focus of the current study took into account newspaper articles written about two major candidates running for president of Uruguay in the recently held national elections: Daniel Martínez and Luis Lacalle Pou, respectively. These politicians were leading the polls of public opinion  and represented political parties of contrasting ideologies (left and conservative, respectively). Newspaper articles written about Daniel Martinez and Luis Lacalle Pou dated to the time when primary elections were held in Uruguay last June 30th of 2019, and these politicians were elected to represent their political parties at the national elections on October and November of 2019.
The newspapers considered for this study were the following:
Primary Elections Period (June 1 to July 25 of 2019) - 440 news articles total:
El Observador (print & digital) https://www.elobservador.com.uy
La Repùblica (print & digital) https://www.republica.com.uy
Montevideo Portal (digital only) https://www.montevideo.com.uy/index.html
Political Debates between Daniel Martínez and Luis Lacalle Pou (October, 1 and November, 13 of 2019) - 90 news articles total:
La Diaria (print & digital) https://ladiaria.com.uy
La Red21 (digital only) http://www.lr21.com.uy
The main corpus used as a dataset needed to train machine learning algorithms was derived from a total of 440 articles published by El Observador, La Repùblica, and Montevideo Portal in order to analyze document similarities (clustering) and to predict topics (topic modeling). Another 90 news articles published La Diaria and La Red21 in addition to El Observador, La República and Montevideo Portal on the political debate between the two political candidates previously mentioned were used instead to predict the political leaning for each newspaper based on the word context for a selected number of topics.
AUTOMATED DETECTION OF POLITICAL IDEOLOGY FROM TEXT_
In this study the author focused on news articles published by newspapers, which for many persons is the primary source of information and thus; they play a pivotal role in shaping personal and public opinion. Automatic detection of political ideology from news articles is based on the notion that journalists can modulate the reader's perception of a political topic through word choice; for instance when the author employs word usage with positive or negative connotations when referring to a political candidate or party, or by varying the credibility of the source . The ideological perspective of a journalist is also often expressed in the choice of discussed topics as journalists with opposing ideologies will choose to write on different topics and make them more salient according to their views. Nonetheless, newspapers don't explicitly express their political preferences, which makes the task of detecting political ideology in news articles somewhat difficult.
Machine learning in conjunction with natural language processing algorithms have been implemented for the automated detection of political ideology from text. For instance, Elfardy et al, identified the ideological perspective of a person by using semantic features derived from the person's written texts . The use of machine learning algorithms for detection of political ideology from news articles was also explored by Kulkarni and colleagues . They proposed a model (based on Bayesian approach with stochastic attention units) that leveraged not only the text contained within news articles but also their titles and hyperlink structure (news articles would provide weblinks to other media sources with similar political ideology) as means to rank 59 news sources based on their predicted political ideology . Gentzkow and Shapiro instead constructed an index of media's political bias that measured the similarity of news outlet's language to that of a congressional Republican or Democrat according to written text derived from Congressional Records in 2005 . Their index measured the frequency of language usage that would 'sway' readers to the left or to the right on political issues; by examining the set of all phrases used by members of Congress during 2005, and identifying those phrases that were more frequently used by Democrats or Republicans. Consequently, they indexed newspapers by the degree in which they used 'politically charged' phrases in their news articles that were reminiscent of those used in political speeches by Democrat or Republican politicians . The resulting index allowed the authors to compare newspapers to one another, rather than comparing them to any given standard of 'true' or 'unbiased' journalism . Iyyer et al, implemented a recursive neural network for detection of ideological bias at the sentence level , differing from previous approaches that were based on 'bag of words' classifiers. The authors were interested in learning representations that could distinguish political bias given labeled data; with their dataset derived also from Congressional Debates during the year of 2005 . Ahmed and Xing  developed topic models (multi-view Latent Dirichlet Allocation) capable of recognizing the ideological bias in a given document, their approach was also capable of summarizing where the bias was manifested on a topical level, and provided readers with alternate views that would help them to remain informed from different perspectives. Lazaridoul and Krestel focused on the examination of 'selection bias' in the sense of how much space a newspaper dedicates for each political party; they also examined how often politicians were mentioned and how often politicians were quoted in news articles .
Despite all the technical advances previously mentioned, a key question still remains to be answered and is: How the automated detection of ideological bias in news articles will eventually contribute to unbiased journalism and a more balanced coverage of political events and social issues to news readers in Uruguay? The work discussed here at least help towards this goal by providing the Uruguayan reader with an objective analysis of political journalism and its inherent bias so as to promote critical thinking and prevent the reader from potential manipulation by the written press.
TECHNICAL IMPLEMENTATION & RESULTS_
The corpus was assembled by retrieving articles containing the name of the political candidate (either Luis Lacalle Pou or Daniel Martínez) by typing their names in the search box within the website of three newspapers: El Observador, Montevideo Portal, La República; whereas news articles on both political debates were downloaded from the website of newspapers during the same or the following days to the debates. The Python library NLTK  was used to estimate the size of the corpus in terms of tokens and is shown on Table 1.
Table 1. Structure of corpus used in this study in terms of number of news articles and document size in tokens for each newspaper and document of proposal for governance.
It is important to note that because the search of articles for each political candidate was conducted independently even within the same newspaper, news articles mentioning both candidates are thus shared between files: that is the same text file mentioning Luis Lacalle Pou and Daniel Martínez is shared between collections (a) El_Observador/Luis_Lacalle_Pou and (b) El_Observador/Daniel_Martínez.
The conjunction of news articles searched for 'Daniel Martínez' contained more tokens relative to the conjunction of news articles searched for 'Luis Lacalle Pou' for El Observador and Montevideo Portal newspapers but not for La República. During the time period of the study (June-1 to July-25 of 2019) La República had a considerably minor journalistic output in terms of published news articles, and consequently tokens, relative to the other two newspapers.
As part of the corpus, the author also included the programatic lines / proposals for governance for each of the political candidates [25, 26]. The proposals for governance were used as reference to compare word contexts of selected topics from news articles on political debates published by the five newspapers shown on Table 1.
Examining mentions of political pre-candidates from news articles_
When news articles are assembled together into a single text file according to their published date, the location of words of interest along a temporal line can be determined. This positional information can be displayed using dispersion plots (Figure 1 and 2) and served the purpose of investigating changes in language use over the studied time period (June 1 to July 2015 with preliminaries held on June 30 of 2019).
Figure 1. Dispersion plots for news articles searched and downloaded using the keyword 'Luis Lacalle Pou' from El Observador (top), Montevideo Portal (middle), and La República (bottom) newspapers, respectively. Each blue mark represents an instance of a word, whereas each row represents the entire corpus composed of news articles arranged into a single document according to publication date from June 1 (left) to July 25 (right) of 2019.
Figure 2. Dispersion plots for news articles searched and downloaded using the keyword 'Daniel Martínez' from El Observador (top), Montevideo Portal (middle), and La República (bottom) newspapers, respectively. Each blue mark represents an instance of a word, whereas each row represents the entire corpus composed of news articles arranged into a single document according to publication date from June 1 (left) to July 25 (right) of 2019.
From Figures 1 and 2 it can be seen that the frequency of mentions for the leading candidates Luis Lacalle Pou and Daniel Martínez was higher across the entire time period relative to their competitors within the same party (Juan Sartori, Jorge Larañaga, Enrique Antía, and Carlos Iafigliola for Partido Nacional; and Carolina Cosse, Mario Bergara, and Óscar Andrade for Frente Amplio; respectively). Luis Lacalle Pou and Daniel Martínez eventually won their primaries on June 30 of 2019 and both politicians selected running mates (candidates for Vice President) soon afterwards (Beatriz Argimón was selected by Luis Lacalle Pou, and Graciela Villar was selected by Daniel Martínez, respectively). The appearance in mentions for Argimón and Villar are clearly seen on the dispersion plots shown, specially for Villar whose selection as running mate of Martínez was considered controversial because of her status as outsider of the mainstream political system, in addition to the controversy surrounding her academic degree and title credentials [27, 28].
The percentage of mentions for each politician relative to the size of the news article's corpus for each newspaper is shown on Figure 3. The frequency of politician's mentions in news articles corresponded well with polls results after the election for each party .
Figure 3. Percentage of politician's mentions relative to corpus size for each newspaper. Because in many news articles Luis Lacalle Pou is referred as 'Lacalle' or 'Luis Lacalle', the author included the percentage of independent mentions for both last names, 'Lacalle' and 'Pou', respectively. Cross-referencing of mentions for political rivals was also included; for instance the percentage of mentions of Daniel Martínez in news articles searched and downloaded with Luis Lacalle Pou as keyword for each newspaper, and the percentage of mentions for Lacalle and Pou in news articles searched and downloaded with Daniel Martínez as keyword for each newspaper.
The results shown on Figure 3 revealed that for La República, the nomination of Graciela Villar as candidate for vice-president for Frente Amplio was very relevant; and differed from Montevideo Portal and El Observador whose emphasis was on Carolina Cosse who was second in votes after Martínez but was not chosen to run on the ticket as vice-president along Martínez. On the other side, it can be seen that the chosen running mate for Luis Lacalle Pou, Beatriz Argimón, did not generate media buzz as her mentions were proportionally less than those to other candidates such as Sartori who was a newcomer and obtained a second place in the primaries.
Identification of the predominant political party in news articles_
In order to investigate the role of Partido Nacional and Frente Amplio in the content of news articles from El Observador, Montevideo Portal, and La República newspapers, the author created a frequency-based score of Partido Nacional and Frente Amplio based on certain words that reflected each party:
Partido Nacional >
'lacalle', 'larrañaga', 'antía', 'sartori', 'nacionalista', 'blanco', 'pn', 'oposición', 'argimón'
Frente Amplio >
'martínez', 'cosse', 'bergara', 'andrade', 'frente', 'frentista', 'frenteamplista', 'fa', 'oficialismo', 'villar'
The author then analyzed the presence of these words for each sentence in news articles for each newspaper in order to classify sentences into four categories: (a) sentences assigned to Partido Nacional; (b) sentences assigned to Frente Amplio; (c) sentences that contained words descriptors for both political parties; and (d) sentences without any references to political party's words descriptors, and thus of unknown assignment. The results are shown on Figure 4 and depicted that the biggest percentage of political party associated sentences were found on news articles from La República newspaper, indicating possible ideological bias, specially when sentences associated with Partido Nacional were completely absent from news articles searched and downloaded using the keyword Daniel Martínez. The author also looked at cross mentions, this means sentences assigned to Frente Amplio when news articles were searched and downloaded using Luis Lacalle Pou as keyword; and similarly when sentences were assigned to Partido Nacional when news articles were searched and downloaded using Daniel Martínez as keyword. Cross mentioning measures the balance in political writing since no politician is consistently reported in isolation from his political context and thus from his political rivals. For instance, cross mentions to Frente Amplio in news articles searched and downloaded on Luis Lacalle Pou as keyword had slightly over 9% of sentences assigned; whereas cross mentions to Partido Nacional in news articles searched and downloaded using Daniel Martínez as keyword had 5.7 and 6.0% of sentences assigned to, with La República not having any mention at all towards Partido Nacional for articles searched and downloaded when Daniel Mart[inez was used as keyword. This phenomena was also evident when sentence assignation to political party affiliation was calculated for each day during the period of study (Figure 5).
Figure 4. Percentage of sentences (relative to total sentences in text file) assigned to political affiliation (Partido Nacional, Frente Amplio or both) from news articles searched and downloaded from El Observador, Montevideo Portal, and La República, respectively. Unassigned sentences are shown in yellow. Results are based on combined news articles into a single file during the time period of study (six files in total: two files (Lacalle Pou / Martínez) for each newspaper).
Figure 5. Daily percentage of sentences assigned to political affiliation (Partido Nacional, Frente Amplio or both) from news articles searched and downloaded from El Observador, Montevideo Portal, and La República respectively. News articles published during the same day for each newspaper were combined into a single text file and analyzed in conjunction. Unassigned sentences are not shown. Primaries were held on June 30 of 2019.
From Figure 5 it can be seen that the amount of news days across the three newspapers in which cross mentioning of Frente Amplio exceeded 10% of sentences when articles were searched and downloaded using Luis Lacalle Pou as keyword was superior; compared to the amount of news days in which cross mentioning of Partido Nacional exceeded 10% of sentences when articles were searched and downloaded using Daniel Martínez as keyword. This can be explained by the combination of two processes: first, the primary role of Partido Nacional as the leading party of the opposition towards Frente Amplio and the current government establishes that their media appearances contain a rhetoric of criticism towards Frente Amplio and thus much more mentions which get published by newspapers because the primary type of journalism in Uruguay is 'declarative' in nature. Second, Frente Amplio as the governing party does not necessarily need to mention Partido Nacional when campaigning, as they need to convince the electorate to continue in power for the next cycle. A similar phenomena of superior mentions in media towards the ruling party was also observed in England .
The assignment of sentences to political affiliation was also studied by the author when news articles were published reporting on the televised political debates between Luis Lacalle Pou and Daniel Martínez held on the nights of October 1 and November 13 of 2019 (Figure 6). The results showed that sentences assigned to both political parties (Partido Nacional and Frente Amplio mentioned on the same sentence) were predominant in news articles published by the newspapers El Observador, Montevideo Portal, and La República; whereas sentences assigned to Frente Amplio alone were the predominant ones in newspapers such as La Diaria and La Red21, respectively. The journalistic reporting from La República was definitely different when reporting on the political debate compared to the reporting of the political campaign during the time period studied (compare Figure 4, 5 and 6). There was a clear drop in sentences containing mentions to both political parties in La Diaria and La Red21 newspapers relative to the other three media outlets. Similar to previous results shown on Figures 4 and 5, Frente Amplio was the most mentioned political party when reporting on the debates.
Figure 6. Percentage of sentences assigned to political affiliation (Partido Nacional, Frente Amplio or both) from news articles searched and downloaded on the political debates between Luis Lacalle Pou and Daniel Martínez held on October 1 and November 13 of 2019. News articles for each newspaper were combined into a single text file (5 text files, one for each newspaper) and sentences parsed according to frequency counts for party-related words in them. Words considered as party-related were as follow: Partido Nacional > 'lacalle', 'larrañaga', 'antía', 'sartori', 'blanco', 'blancos', 'nacionalista', 'pn', 'oposición', 'argimón', 'abdala', 'delgado', 'alternancia', 'posadas'; Frente Amplio > 'martínez', 'cosse', 'bergara', 'andrade', 'frente', 'frentista', 'frenteamplista', 'fa', 'oficialismo', 'oficialista', 'villar', 'miranda', 'mujica', 'astori', 'vázquez', 'bonomi', 'leal', 'sendic', 'socialista'. The number of news articles sampled from each newspaper were: El Observador (30), Montevideo Portal (15), La República (26), La Diaria (14), La Red21(5), respectively.
Abstracting the content (core topics) from news articles_
Unsupervised machine learning techniques such as text clustering and topic modeling helped the author to automate the elucidation of content within news articles and thus asses their relationship based on similarity in their word usage [30, 31, 32]. News articles that were similar to each other could be grouped/clustered together and the resulting clusters generally portrayed the overarching topics, themes and/or patterns that related to the newspaper's vision of the political scene it was reporting at the time. When topics were discrete and well discerned, clusters did not overlap. On the contrary, when topics were fuzzy, documents were hard to distinguish and resulted with overlap among clusters. Clustering of news articles per newspaper was used to determine the number of clusters and thus the number of topics that were discussed during the time period of study (Figure 7 and Figure 8). The Python code used is available upon request.
Figure 7. Results from k-Means clustering algorithm (MiniBatchKMeans implementation) for news articles searched and downloaded using 'Luis Lacalle Pou' as keyword. Left column shows clustering of 102 news articles retrieved from 'Montevideo Portal' newspaper. Center column shows clustering of 88 news articles retrieved from 'El Observador' newspaper. Right column shows clustering of 30 news articles retrieved from 'La República' newspaper. First two rows of graphs display dimensionality reductions of vector space into 2D scatter plots as 'Principal Component Analysis' (PCA) (first row) and 't-distributed stochastic neighbor embedding visualization' (t-SNE) (second row). Silhouette Scores (third row) and Elbow Curves (fourth row) are shown as menas to evaluate the results of selecting the number of optimal k-clusters. The Python machine learning visualization library Yellowbrick  was used to generate t-SNE, Silhouette Scores and Elbow Curves, respectively.
Figure 8. Results from k-Means clustering algorithm (MiniBatchKMeans implementation) for news articles searched and downloaded using 'Daniel Martínez' as keyword. Left column shows clustering of 121 news articles retrieved from 'Montevideo Portal' newspaper. Center column shows clustering of 94 news articles retrieved from 'El Observador' newspaper. Because only 5 news articles were retrieved from 'La República' newspaper, no clustering study was implemented and instead, two clusters (and thus two topics) were arbitrarily assigned. First two rows of graphs display dimensionality reductions of vector space into 2D scatter plots as 'Principal Component Analysis' (PCA) (first row) and 't-distributed stochastic neighbor embedding visualization' (t-SNE) (second row). Silhouette Scores (third row) and Elbow Curves (fourth row) are shown as menas to evaluate the results of selecting the number of optimal k-clusters. The Python machine learning visualization library Yellowbrick  was used to generate t-SNE, Silhouette Scores and Elbow Curves, respectively.
From the above figures (Figures 7 and 8) it can be seen that news articles published by 'Montevideo Portal' were more diverse in terms of their content (clustered into 6 groups for both Luis Lacalle Pou and Daniel Martínez) relative to articles published by 'El Observador' (clustered into 5 and 4 groups for Luis Lacalle Pou and Daniel Martínez respectively). Articles published by 'La Reúplica' were less in number (30 articles retrieved when 'Luis Lacalle Pou' was used as keyword; and only 5 articles retrieved when 'Daniel Martínez' was used as keyword) and for the case of 'Luis Lacalle Pou', much more diverse in content which resulted into a high number of clusters (11 clusters). From this analysis (text clustering), the author proceeded to implement topic modeling as means to identify the main topics being discussed by each daily newspaper. For each topic, the 100 most frequently mentioned terms were retrieved and from these, the most interesting ones deemed by the author are shown. Notice that the term associated with 'Luis Lacalle Pou' does not occur within any of the topics; this is so because text clustering and modeling algorithms discard any term that appears in more than 85% of the news articles (discard too common terms), which is the case because all articles contain Luis Lacalle Pou term because they were retrieved and downloaded using this keyword. The same is for Daniel Martínez. The clustering and topic modeling algorithms also discarded terms that occurred in less than 5% of the articles. The topics obtained were the following and their relationship according to sharing terms within topics is shown on Figure 9 and Figure 10:
El Observador > Luis Lacalle Pou > 5 Topics (OLLPT1-T5)
Sartori - Larrañaga - nacional - empresario - Argimón - directorio - política - precandidato - campaña - interna - Antía - denuncia - alianza - acto - precandidatos - nacionalista - senado - presidenta - fórmula - intendente - sector - diputado - Costa - líder - blancos - dirigentes - candidato - falsas - Moreira - Herrera - Abdala - vicepresidente - senador - Wilson - blanco - vivir - mujer - político - blanca - embargo - país
Talvi - Martínez - Colorado - Cosse - voto - encuestas - encuesta - Sanguinetti - Radar - intención - Opción - Andrade - interna - elecciones - internas - Equipos - datos - votar - resultados - Factum - Larrañaga - ecuestadoras - opinión - Bergara - Cifra - canal - números - margen - error - nacional - consultora - partidos - votos - candidatos - votantes - indecisos - pública - electoral - puntos cabildo - consultores - informe - Sartori - liderazgo - electorado - precandidatos - expresidente - ventaja - votación - políticos - economista
programa - propuestas - Silveira - gobierno - ley - seguridad - propuesta - educación - nacional - gasto - equipo - candidato - país - social - técnicos - sectores - estado - trabajo - políticas - ahorrar - reforma - blanco - estatal - internacionales - sociales - desarrollo Iafigliola - públicas - déficit - medicamentos - inversión - convención - laborales - sector - negociaciones - descentralización - Equipos - senador - sistema - mayoría - acuerdo - parlamento - aborto - economista
gobierno - proceso - precandidato - partidos - rumbo - país - campaña - gira - acto - internas - multicolor - gobernar - presidente - etapa - nacionalista - militancia - social - elecciones - militantes - sector - actos - república - discursos - Vázquez - candidatos
Frente - Amplio - Talvi - Martínez - Villar - fórmula - izquierda - candidato - Manini - votos - oposición - colorados - centro - coalición - gobierno - elección - presidencial - política - Astori - blancos - debate - candidatos -derecha -victoria -colorado - opositor - líder - Batlle - elecciones - campaña - votación - oficialismo - interna - Cabildo - votantes - economía - historia
Montevideo Portal > Luis Lacalle Pou > 6 Topics (MPLLPT1-T6)
partido - candidato - fórmula - interna - Argimón - nacional - Sartori - elección - reforma - vivir - Martínez - programa - campaña - propuestas - política - seguridad - mujer - colorado - candidatos - directorio - Talvi - Larrañaga - ideas - educación - propuesta - miedo - vicepresidente - precandidato - acuerdo - diálogo - nacionalista - condiciones - vicepresidencia
encuesta - voto - Sartori - Antía - Larrañaga - Radar - partido - interna - nacional - votantes - blanca - consultora - candidato - Opción - votar - elecciones - internas - amplio - frente - colorado - encuestas - votará - crecimiento - lidera - competencia - Manini - Montevideo - precandidato - votos - vtv - elección
Talvi - Martínez - votos - Andrade - Cosse - partido - frente - amplio - escrutinio - Sanguinetti - colorado - Bergara - corte - Amorín - Batlle - resultados - internas - electoral - abierto - cabildo - precandidatos - encuestadoras - Larrañaga - partidos - nacional - datos -Equipos - Antía - Sartori - votantes - Manini - elecciones - frenteamplista - candidatos - Iafigliola - votará - interna - debatir
información - ministerio - asesor - social - sociales - trabajadores - declaraciones - desarrollo - programa - educación - políticas - gente - pública - educativo - entrevista - trabajo - público - propuestas - datos - gobierno - Montevideo - denuncia - seguridad - ley - propuesta
lista - falsas - noticias - Sartori - denuncia - campaña - redes - equipo - periodista - precandidatos - programa - precandidato - denuncias - sociales - democracia - políticos - listas - dirigentes - Montevideo - radio - ciudad - electoral - nacional - blanco - tema - compañeros
gobierno - país - uruguayos - presidente - precandidato - nacionalista - senador - líder - campaña - pueblo - ministro - política - gobernar - sector - uruguayo - oposición - blanco - partido - votar - Venezuela - acto - discurso - economía - estado - justicia - nacional - república
La República > Luis Lacalle Pou > 11 Topics (RLLPT1-T11)
interna - redes - gobierno - resultado - blanca - democracia - elección - denuncia - Sartori - senador - precandidato - fórmula - precandidatos - programa - libertad - uruguay - escenario - periodista - electoral - sociales - acuerdo - relación - historia - directorio
Opción - elecciones - Radar - nacionalistas - fa - electorado - encuesta - candidato - Talvi - colorados - consultores - fidelidad - Sanguinetti - frenteamplista - internas - Larrañaga - Sartori - nacionales - colorado - Martínez - amplio - frente - evidencia - filas - votarán - Antía - presidencial - votantes - votaron - Batlle - candidatos - cabildo - abierto - partidos - interior
coalición - partidos - gobierno - Talvi - proyecto - Silveira - abierto - cabildo - independiente - colorado - propuestas - candidato - alianza - gobernar - programa - Manini - país - política - gente - Mieres - multicolor - debates - acuerdos - blanco - campaña - elecciones - competir - cooperar - técnicos - equipos - militantes - programáticas - debate - convicción - entrevista - presidente
Herrera - senado - precandidatura - expresidente - convención - gobierno - dirigentes - parlamento - información - interna - político acto - gobernar - embargo - uruguayos - candidatos - ideas - república - independiente - presidente
Sartori - campaña - diario - empresario - Larrañaga - Antía - información - programa - precandidato - internas - precandidatos - sector - pueblo - blancos - ley - sede - acuerdo - encuestas - nacionales - senador - corte - electoral - comando
candidata - fórmula - Argimón - vicepresidente - vicepresidenta - compañera - blanca - compañeros - equipo - mujeres - presidenta negociadora - técnico - mujer - sector - capacidad - elección - parlamento - candidato - vicepresidencia - directorio - político - jurídico diario - precandidatos - actos
edad - jubilatoria - jubilarse - trabajar - económicamente - jubilen - estimule - contrato - trabajadores - coalición - trabajando - senador reforma - seguridad - dirigentes - programa - social - miedo - blanco - subrayado - constitución - sistema - plebiscito - partidos - reglas vivir - laboral - rumbo - declaraciones - gobierno - fuentes - propuesta - programas - presidenta - ciudadanía
precandidato - red - ministerio - asesor - políticas - presidencia - sociales - interna - fórmula - denuncia - debatirá - desarrollo - debatir - candidato - social - vicepresidente - programa - senador - nacionalista - internas - directorio - edad - seguridad - convención - jubilatoria - constitución - sistema - emprendimiento - vivir - subrayado
gobierno - emprendimiento - evolucionar - país - upm - tarifas - publicidad - uruguayos - electoral - campaña - amplio - frente - gobiernos - bonanza - presidente - presidencia - celulosa - república - intereses - ejecutivo - candidato - discurso - económico - económica - acto
debatir - Andrade - debate - Cosse - Talvi - precandidatos - Sartori - debates - nacionalistas - frenteamplistas - semanario - búsqueda canal - radio - uruguay - internas - partidos - frente - amplio - acto - convención - vicepresidente - candidata - precandidato - consultora - entrevistado
Martínez - Talvi - amplio - colorado - frente - Bergara - Batlle - consultores - votos - Sanguinetti - Andrade - votaría - datos - Cosse - balotaje - candidato - Opción - votantes - equipo - escenario - Larrañaga - elecciones - Radar - Antía - internas - Sartori - encuesta electo - precandidatos - blancos - encuestadora
Figure 9. Network graph visualization of the relationship between topics (and their shared terms) for each of the three newspapers studied. News articles retrieved and downloaded using 'Luis Lacalle Pou' as keyword were subjected to topic modeling and the associated terms for each topic for each newspaper were visualized in a network graph using Gephi . Topics formed 8 'communities' (color coded) based on the degree of shared terms in which La República newspaper shared more topics with Montevideo Portal newspaper (5 communities) compared to topics shared with El Observador newspaper (3 communities).
El Observador > Daniel Martínez > 4 Topics (ODMT1-T4)
Cosse - fórmula - MPP - candidato - mujer - Mujica - fa - votos - lista - exministra - interna - Andrade - industria - Vázquez - presidencial - oficialismo - expresidente - política - dirigentes - fuerza - exprecandidata - campaña - vicepresidenta - Lustemberg - comunista - diputada - sector - paritaria - elecciones - votada - candidatura - candidata - Miranda - presidente - partido - senadora
estado - gestión - gobierno - fiscal - URSS - precandidato - izquierda - políticas - ciudad - política - gasto - intendencia - público - Montevideo - seguridad - IMM - austeridad - país - Venezuela - socialista - dictadura - derechos - déficit - mundo - oposición - departamento - campaña - sistema - impuestos - democracia - visión - propuesta - reforma - públicas - humanos
Villar - social - fórmula - asamblea - exedila - candidata - vicepresidenta - Uruguay - departamental - Astori - junta - plenario - candidato - título - fa - militante - psicología - elección - proceso - dirigente - Topolansky - sector - Observador - Seregni - referentes - Cosse - derecha - militantes - lista - huella - Montevideo - Sendic - frenteamplista - Bergara - dirigentes - astorismo - sindical - ministro - formación - Manini - oficialismo - campaña
Talvi - partido - Lacalle - Pou - colorado - Sanguinetti - nacional - candidatos - voto - Andrade - Sartori - internas - debate - interna - elecciones - Larrañaga - intención - votantes - encuestas - Radar - votos - partidos - blanco - blancos - votación - Equipos - debatir - resultados - colorados - bloque - reforma - encuesta - Factum - nacionales - escenario - electoral - precandidato - Cifra - Bergara - mayoría - izquierda - elección - victoria - canal - candidato - porcentaje - campaña - precandidatos - opción - opositor
Montevideo Portal > Daniel Martínez > 6 Topics (MPDMT1-T6)
precandidato - fórmula - Uruguay - reforma - país - campaña - seguridad - gobierno - precandidatos - frenteamplista - candidato - nacional - trabajo - programa - presidente - Cosse - votar - declaraciones - fa - Larrañaga - blanca - desarrollo - radio - interna - paritaria - Astori
Talvi - Lacalle - Pou - Sanguinetti - partido - Sartori - Larrañaga - Batlle - escrutinio - Antía - colorado - Andrade - resultados - ganador - Carolina - nacional - Bergara - internas - Cosse - Factum - precandidatos - votos - opción - interna - elecciones - consultora - electoral - liceo - ventaja - colorados - cifras
Villar - fórmula - compañera - candidata - Cosse - social - política - vicepresidenta - título - asamblea - vicepresidencia - Montevideo - junta - edila - pueblo - departamental - candidato - elecciones - mujer - Vázquez - Uruguay - Sendic - movimiento - presidencial - precandidata - presidencia - sector - plenario - sindical - prensa - conferencia - campaña - nacionales - fa - capacidad - sociales - dirigente - militante - salud - presidente - Seregni - exedila - huella - Bergara - cargo
votos - cabildo - Lacalle - abierto - Pou - Talvi - partidos - tendencia - partido - voto - Manini - electoral - interna - políticas - votó - crecimiento - datos - cifras - candidato - escrutinio - votantes - nacional - votación - precandidatos - candidatos - colorado - elección - acuerdos - internas - elecciones - Radar - oposición - Sartori - pueblo - votada - grupo - Mujica - análisis - Sanguinetti - derecha - estado - independiente - gobierno - discurso - Miranda - nacionales - política
Amado - debates - diputado - debate - ley - proyecto - candidatos - Montevideo - partido - propuesta - elecciones - fa - Talvi - izquierda - resultados - presidencia - elección - votación - nacionales - colorado - candidato - partidos - colorados - ejecutivo trabajadores - sector - constitución - Mieres - frentista - cámara - independiente - diálogo - cambios - político - ciudadanos - blancos - presidenciales - presidente
voto - intención - encuesta - Andrade - Cosse - puntos - Radar - partido - estudio - interna - Bergara - votantes - nacional - opción - Factum - colorado - consultora - técnico - votar - análisis - ventaja - lidera - internas - líder - candidatos - crecimiento - encuestas
La República > Daniel Martínez > 2 Topics (RDMT1-T2)
país - fórmula - cárceles - Villar - plenario - unidad - nacional - desarrollo - justicia - Uruguay - compromiso - partidos - debate - política - delitos - sectores - paritaria - candidato - fa - precandidato - transformaciones - transformar - renovación - oportunidades - militancia - presos - gobierno - social - iniciativa - posibilidad - integración - avances - presidente - candidata
votar - interna - frenteamplistas - votantes - Cifra - Cosse - encuestador - encuesta - pública - votarán - partidarias - acuerdo - respaldo - opinión - precandidatos - Bergara - internas - candidato - Villar - plenario - nacional - sectores - mejor - candidata - presidente - presidencial - representantes - vicepresidenta - república - vicepresidencia - Montevideo - compromiso
Figure 10. Network graph visualization of the relationship between topics (and their shared terms) for each of the three newspapers studied. News articles retrieved and downloaded using 'Daniel Martínez' as keyword were subjected to topic modeling and the associated terms for each topic for each newspaper were visualized in a network graph using Gephi . Topics formed 7 'communities' (color coded) based on the degree of shared terms. La República newspaper shared a topic with Montevideo Portal newspaper (one community) in relation to El Observador newspaper, with no communities shared.
Document clustering, and topic modeling in conjunction with network graph visualization shed light on the closeness of topic selection among news articles from Montevideo Portal and La República in relation to El Observador newspaper; at least when community structure was investigated through network graph visualization (Figure 9 and Figure 10). It is interesting to note as well that the nodes derived from topics discussed by El Observador were more dispersed in both network graphs, as opposed to node aggregation from topics discussed by Montevideo Portal. These results were in accord with results obtained through clustering analysis (PCA and T-SNE results shown on Figure 7 and Figure 8). From all this, the author concluded that news articles from El Observador were more diverse in content and topic selection for both, Luis Lacalle Pou and Daniel Martínez related news coverage.
Comparative analysis of contextual meaning from news articles relative to proposals for governance documents_
Next, the author identified 130 topics terms that were shared between the three newspapers (Figure 9 and Figure 10 combined). These terms really represented the conversation that took place before, during and after the primary election of June 30 of 2019; and displayed the breath of topics discussed by newspapers. The topics were the following:
acto - acuerdo - amplio - Andrade - Antía - Argimón - asamblea - Astori - Batlle - Bergara - blanca - blanco - blancos - cabildo -campaña - candidata - candidato - candidatos - ciudad - colorado - colorados - consultora - Cosse - datos - debate - debatir - democracia - denuncia - departamental - derecha - desarrollo - diputado - directorio - dirigente - dirigentes - economía - educación - elecciones - elección - electoral - encuesta - encuestas - equipo - equipos - estado - exedila - fa - factum - falsas - frente - frenteamplista - fórmula - gobernar - gobierno - huella - Iafigliola - intención - interna - internas - izquierda - junta - Lacalle - Larrañaga - ley - lista - líder - Manini - Martínez - militante - ministro - Miranda - Montevideo - mujer - mujeres - Mujica - nacional - nacionales - nacionalista - opción - oposición - paritaria - partido - partidos - país - plenario - política - políticas - político - políticos - Pou - precandidato - precandidatos - presidencial - presidente - programa - propuesta - propuestas - puntos - pública - público - radar - reforma - república - resultados - Sanguinetti - Sartori - sector - seguridad - senador - Sendic - Seregni - sindical - social - sociales - Talvi - trabajo - título - Uruguay - Vázquez - Venezuela - ventaja - vicepresidenta - vicepresidente - Villar - vivir - votación - votada - votantes - votar - voto - votos
From these, the author selected 14 terms (written in orange color above) that he considered relevant enough to compare the context (and thus their meaning) in which they were written in news articles from the political debates relative to the proposals for governance for each political party. By analyzing and comparing these words in their native contexts, the author hypothesized that contexts from news articles may be similar to contexts in proposal for governance documents if the newspaper in question is ideologically biased towards one political party or the other. With this in mind, one could assess political ideology by plotting the word context for each of the 14 terms previously mentioned into a network graph and subsequently identifying the resulting community structure of the graph. For instance, if nodes for word contexts from El Observador newspaper and nodes for word contexts derived from Partido Nacional's proposal for governance document share more communities in common than with word contexts from Frente Amplio proposal for governance, one could conclude that El Observador displays ideological bias towards Partido Nacional. The same approach stands true for ideological bias towards Frente Amplio. This approach assumed that sharing of ideological views was expressed through similar word usage.
In order to extract from text the selected 14 terms along with their word contexts, a concordance analysis using Python's NLTK library was used. Concordance analysis were previously used by the author to analyze the word contexts from his personal Facebook's messages exchanges with friends . For instance, let's see the word context for the term 'ley'/'law' for each document of governance (Partido Nacional and Frente Amplio respectively) in which there were clear differences:
Propusta de Gobierno Partido Nacional - Proposal for Gobernance Partido Nacional
Displaying 30 of 119 matches:
tarios y solicitudes formuladas al amparo de la Ley N° 18.381, de Acceso a la Información Pública
́n. Nos proponemos mejorar lo ya previsto en la ley reguladora del estatuto del funcionario de la A
an se realizarán a través de los proyectos de ley de urgente consideración y de presupuesto quin
sistema de acceso a la información pública La Ley de Acceso a la Información Pública fue un ava
Acceso a la Información Pública creada por la Ley N°18.381, para los casos en que las empresas pu
r único”, en el marco de lo establecido por la Ley N° 16.832 (Ley de Marco Regulatorio). Esto le p
n el primer semestre de gobierno un proyecto de ley que incluya los siguientes puntos: Fortalecer a
r único”, en el marco de lo establecido por la Ley N° 16.832 (Ley de Marco Regulatorio). Con el fi
testad del ente, sino que debe ser decidido por ley y con recursos contabilizados en forma separada
arla a una realidad cambiante. Elaboraremos una Ley de Telecomunicaciones y Servicios Digitales que
sto se logrará mediante la elaboración de una ley específica sobre desechos de aparatos eléctri
la redacción, aprobación y aplicación de una Ley de Transporte que, junto con sus reglamentacion
do de Fomento de Marina Mercante, creada por la ley N°14.650, en un órgano consultivo preceptivo e
rimer ejemplo es lo ocurrido a propósito de la ley de negociación colectiva del sector privado (L
ceder al mecanismo de descuelgue previsto en la ley de negociación colectiva. 4. La defensa del po
ar las observaciones efectuadas por la OIT a la ley de negociación colectiva. Es hora de que una l
y de negociación colectiva. Es hora de que una ley clave del sistema de relaciones laborales cumpl
con los apoyos más amplios posibles, una nueva Ley de Seguridad y Salud en el Trabajo que se adapt
en el sistema educativo. 11. Aprobación de una Ley de Acoso Moral en los lugares de trabajo, que d
o, etc.). La tercera es elaborar un proyecto de ley que complemente y profundice las medidas adopta
poyo de todo el sistema político se aprobó la ley No 18.407, que puso fin a una legislación frag
mpatibilizar las disposiciones que surgen de la Ley de Procedimiento Policial con las disposiciones
urre con frecuencia. 15. Modificar el Art. 1 la Ley No 15.605, con el fin de devolver a INAC las co
plimiento de las obligaciones consagradas en la Ley No 15. 939 del 28/12/1987, art. 28, referido a
o de la inversión, se impulsarán proyectos de ley que ya han sido presentados por legisladores de
s, la mayor rentabilidad de la agricultura y la Ley de Promoción de Inversiones justificaron esas
timo elemento para entender la situación es la Ley de Riego. Se trata de una norma bien inspirada,
́s encontraban demasiados obstáculos). Pero la ley generó al mismo tiempo preocupaciones y debate
inversiones necesarias. Pesca y acuicultura La ley de ordenamiento madre del sector pesquero data
endo la investigación tal cual lo establece la Ley N° 19.175 y su decreto reglamentario 115/18. 49
Propusta de Gobierno Frente Amplio - Proposal for Gobernance Frente Amplio
Displaying 30 of 145 matches:
ntes, la negociación colectiva obligatoria, la ley de fueros sindicales, la ley que regula las ter
va obligatoria, la ley de fueros sindicales, la ley que regula las tercerizaciones o la Responsabil
de trabajo, la jornada de trabajo de 8 horas es ley y los laudos mínimos son obligatorios. Todos e
dependencia. En 2015, con la aprobación de la Ley N°19.353, se crea el Sistema Nacional Integrado
dentro de los cuales el matrimonio igualitario, ley de salud sexual y reproductiva, la ley de inter
litario, ley de salud sexual y reproductiva, la ley de interrupción voluntaria del embarazo, regul
mativas para las personas afrodescendientes, la ley de reproducción asistida, la ley de cuotas par
endientes, la ley de reproducción asistida, la ley de cuotas para las personas con discapacidad, s
realizarían en ausencia de las políticas. La ley de promoción de inversiones provee un marco ad
ítica de exoneraciones, apuntando a afirmar la Ley de Promoción de Inversiones como herramienta p
do los recursos económicos comprometidos en su ley de creación, con el objetivo final de constitu
ación de los jóvenes en la ESyS. Impulsar una Ley General de la Economía Social y Solidaria que
ión de la Unidad Agroalimentaria, definida por ley de carácter nacional, fortaleciendo a la DIGEG
está garantizado tanto por la normativa (e.g. Ley 18232) como por el rol central que juegan los o
Fondo de Desarrollo de la Pesca, creado por la Ley N° 19.175. El Fondo permitirá, conjuntamente a
n de compras públicas del Estado que prevé la Ley 19.292. Crear un Fondo de Apoyo a la Pesca Arte
su conjunto. Esto debe estar contemplado en una Ley Nacional de Ciencia, Tecnología e Innovación.
l. En ese sentido, resulta central impulsar una ley nacional de ciencia, tecnología e innovación.
vés de la aprobación e implementación de una Ley de Bioseguridad que incluya los nuevos desafío
territorio. Extender la moratoria aprobada por Ley 19.585 hasta que se investigue en profundidad l
deberá avanzar en la lógica establecida en la Ley General de Residuos, considerándolo como un re
tmosférica y promover la reglamentación de la Ley de Contaminación Acústica. Profundizar la inc
y económico creado por el FA en el marco de la Ley de Ordenamiento Territorial y Desarrollo Sosten
enta un avance muy relevante. En particular, la Ley de Ordenamiento Territorial y Desarrollo Sosten
amiento Territorial y Desarrollo Sostenible, la Ley de Política Nacional de Aguas y la Ley de Desc
ble, la Ley de Política Nacional de Aguas y la Ley de Descentralización Política y Participació
o territorial establecidos en el marco de en la ley 18.308 - Ley de Ordenamiento Territorial y Desa
establecidos en el marco de en la ley 18.308 - Ley de Ordenamiento Territorial y Desarrollo Sosten
es territoriales estratégicas previstas por la ley 19.525 (Directrices Nacionales de OTDS), en tan
iorizar el diseño institucional previsto en la Ley para el tratamiento de asuntos o iniciativas co
The results from the concordance analysis showed marked differences in the interest from each political party towards which laws to regulate, promote and create. For instance, Frente Amplio showed interest in those laws pertaining worker's unions, sexual health and abortion; whereas Partido Nacional showed interest in laws related to telecommunications and digital services, access to public information, recycling of electronic waste and profitability of the agricultural sector.
Let's see now what are the word contexts for 'ley/law' for each set of news articles from the newspapers shown on Figure 6, which were written about the political debates between Luis Lacalle Pou and Daniel Martínez held on October 1 and November 13, respectively. All news articles for both debates for each newspaper were combined into a single text file.
El Observador Political Debate: 30 news articles
Displaying 15 of 15 matches:
stas. En tanto, le remarcó a Lacalle Pou que la ley de procedimiento policial ya habilita a pedir l
Partido Nacional obtuvo 765.990 votos, y por la ley de lemas, su candidato Luis Lacalle Herrera fue
parcial. En esencia, se trataba del proyecto de ley más ambicioso para reformar al Estado, desde su
tos. En el referéndum realizado para derogar la ley que anulaba el monopolio de Ancap, el resultado
do político y programático que se plasme en una ley de urgencia que recoja las coincidencias entre
egido por ley, y si algún candidato se opusiera ley de debate "no percibirá la contribución del Est
argo: conocé los detalles del primer debate por ley 11 de noviembre de 2019 El primer debate obliga
tido por cadena nacional y durará dos horas. La ley indica que la duración no puede ser mayor a ese
r cadena nacional y durará una hora y media. La ley indica que la duración no puede ser mayor a dos
ogado encargado de la redacción del proyecto de ley de urgente consideración anunciado por el candi
inuación le preguntó si consideraba un error la ley de humanización carcelaria que habilitó la libe
o. A continuación remarcó que "nadie conoce" la ley de urgencia anunciada por el nacionalista. "Eso
ahorro de los US$ 900 millones y el proyecto de ley de urgente consideración. En tanto, Martínez di
ncas y la incertidumbre respecto al proyecto de ley de urgente consideración. Incluso dejó para el
del documento de la "coalición multicolor" y la ley de urgente consideración ideada por Lacalle Pou
We can see that El Observador has mentioned laws related to 'procedimiento policial' and 'urgente consideración' that were shown to appear in the document for governance of Partido Nacional. However, El Observador did not mention any laws discussed in the document for governance from Frente Amplio.
Montevideo Portal Political Debate: 15 news articles
Displaying 12 of 12 matches:
y la reciente aprobación en este período de la Ley Integral de Residuos. También considera que la
Hace casi diez meses que el gobierno tiene una ley que autoriza a patrullar las fronteras, pero no
la seguridad. Más estado, más control, pero la ley de la selva no", añadió el candidato del Frente
son las propuestas de Lacalle. Hemos visto una ley de emergencia, más de 300 artículos que nos pre
pedir la cédula en la calle y, sin embargo, la ley de procedimiento policial ya lo permite. Es más
ía". Esa afirmación es verdadera. En efecto, la Ley del Procedimiento Policial habilita en sus artí
del Ministerio del Interior, Gustavo Leal, esta ley se aplica en "operativos policiales" o cuando h
1.11.2019 Este miércoles será el debate que por ley quedó establecido como obligatorio entre los ca
embre de este año se aprobó en el Parlamento la ley 19.827, que "declara de carácter obligatorio" l
3.11.2019 Este miércoles será el debate que por ley quedó establecido como obligatorio entre los ca
embre de este año se aprobó en el Parlamento la ley 19.827, que "declara de carácter obligatorio" l
encia nacional. Tenemos que darle recursos a la ley de violencia de género. Es con la policía pero
We can see that Montevideo Portal has mentioned two laws discussed by two political parties in their documents of governance; and these are 'ley integral de residuos' from Frente Amplio and 'ley de procedimiento policial' from Partido Nacional.
La República Political Debate: 26 news articles
Displaying 13 of 13 matches:
“Nos quedamos sin saber qué van a hacer con la ley de urgencia, con empleados públicos, con Ancap,
. Nos quedamos sin saber qué van a hacer con la ley de urgencia, con empleados públicos, con Ancap,
e todo listo. Esta noche será el debate que por ley quedó establecido como obligatorio entre los ca
embre de este año se aprobó en el Parlamento la ley 19.827, que «declara de carácter obligatorio» l
en propuestas mágicas». DEBATE PRESIDENCIAL «La ley de presunta defensa es como dar un escudo de pl
icó que la propuesta del nacionalista sobre «la ley de presunta defensa, es como dijo un asesor col
estionó si «¿fue una equivocación o un error la ley de liberación de presos? ¿Es un error sostener
con firmeza. Se trató del primer debate que por ley quedó establecido como obligatorio entre los ca
rullaje militar fronterizo para lo cual hay una ley aprobada pero que no se aplica. Propuso declara
la emergencia en la seguridad e insistió con la ley de defensa presunta policial. Martínez exigió q
ida de drogas por sus fronteras y afirmó que la ley de legítima defensa presunta «es un escudo de p
ntales, me da miedo no saber el contenido de la ley de urgencia, eso es lo que nos da miedo, querem
más porque abría la puerta a la historia con la ley que se empieza a aplicar en el ejercicio de la
We can see that La República has primarily mentioned 'ley de urgencia' that is associated with Partido Nacional, the thing is that the context in which it is mentioned is more negative than positive. Because this work did not implement sentiment analysis, the negative sentiment towards Partido Nacional and Luis Lacalle Pou from La República was not assessed.
La Diaria Political Debate: 14 news articles
Displaying 12 of 12 matches:
ntino Mauricio Macri. “Nosotros hemos visto una ley de urgencia que nos preocupa”, sostuvo, y menci
que va a impulsar la economía circular y que la ley de residuos va a hacer que disminuya la cantida
candidato del Partido Nacional no conoce es la Ley Integral de Residuos que se aprobó en este perí
pedir la cédula en la calle y, sin embargo, la ley de procedimiento policial ya lo permite. Es más
ía”. Esa afirmación es verdadera. En efecto, la Ley del Procedimiento Policial habilita en sus artí
del Ministerio del Interior, Gustavo Leal, esta ley se aplica en “operativos policiales” o cuando h
e impulsará una agenda provida sin modificar la Ley de Interrupción Voluntaria del Embarazo, un tem
ttinelli dijo que es posible hacerlo, ya que la ley le da muchas herramientas para que el Ministeri
Le voy a dar otra oportunidad. ¿Fue un error la ley de liberación de presos, que liberó injustament
l candidato oficialista agregó que impulsar “la ley de presunta defensa –como dijo un asesor colora
eclamó a Martínez que “no se esté aplicando” la ley de patrulla de fronteras. El candidato del FA r
ar 900 millones de dólares y el contenido de la ley de urgencia, “que nadie conoce”. Sobre su visió
We can see that similar to La República newspaper, La Diaria also mentioned 'ley de urgencia' with a negative connotation ('nos preocupa', and 'que nadie conoce'). There are mentions to 'ley de procedimiento policial' as in the document of governance from Partido Nacional; and mentions to 'ley integral de residuos', 'ley de residuos' and 'ley de interrupción voluntaria del embarazo' as in the document of governance from Frente Amplio.
La Red21 Political Debate: 5 news articles
Displaying 15 of 15 matches:
andidato frenteamplista dijo que le preocupa la Ley de emergencia del Partido Nacional con artículo
r la cédula de identidad en la calle, cuando la Ley de procedimiento policial ya lo permite. “Le in
es no están preparado para ello. Más Estado, la ley de la selva, no”, remarcó. Desarrollo humano El
circular que este gobierno promoverá junto a la ley de residuos, disminuirá fuertemente la cantidad
gano rector de las elecciones, se indica que la Ley N 19.827 confiere competencia a la Corte Electo
cuerdo a lo establecido en el artículo 2° de la Ley sobre debate obligatorio, el Servicio de Comuni
ndidatos a la Presidencia en cumplimiento de la Ley 19.827, que dispone que es de carácter obligato
tínez y Luis Lacalle Pou, en cumplimiento de la Ley N° 19.827 El mensaje inicial del debate. Lacall
con la ideología que no permite cumplir con la Ley de patrullaje de frontera que es por donde entr
s para combatir el abigeato y dar recursos a la Ley de violencia de género. Con posterioridad, Mart
ra los ricos. Luego preguntó: “¿Fue un error la ley de liberación de presos? sin tener en cuanta si
droga todos los años”, manifestó. Indicó que la ley de presunta defensa es “darle un escudo de plás
, aseguró. En ese momento Lacalle indicó que la ley de patrulla de frontera no se está aplicando, p
da de los uruguayos, o no saber qué contiene la ley de urgencia o que hable de devaluación”, dijo.
tínez y Luis Lacalle Pou, en cumplimiento de la Ley N° 19.827.
We can see that La Red21 negatively mentioned 'ley de urgencia' from Partido Nacional. It also mentioned 'ley de residuos' from Frente Amplio.
To have a more compelling understanding of possible bias in ideology, it is important to analyze the word contexts for the other 13 topic terms mentioned above in the same manner that 'ley/law' was analyzed. The construction of a network graph would help visualize the relationship between shared terms among news articles and the documents of governance, providing a better insight into possible ideological bias in news articles written about the two political debates. The analysis of debate news articles from El Observador in comparison with programs of governance for each political party are shown on Figure 11 and Table 2.
Figure 11. Network graph visualization of the relationship between topics (and their word contexts) for El Observador newspaper and documents for governance from Partido Nacional and Frente Amplio, respectively. About 1,919 nodes and 3,235 edges were visualized using Gephi . Topics were arranged into 12 communities, and each community was color coded. Notice the relationship between nodes (corresponding topics) from El Observador (ObDeb) relative to nodes associated with documents for governance from Partido Nacional (PNprog) and Frente Amplio (FAProg), respectively. News articles written about both political debates (Oct-1 and Nov-13) between Luis Lacalle Pou and Daniel Martínez were used to extract the word contexts associated to 14 topics. The relationship between topics and community structure in the graph is shown more clearly on Table 2 below.
Table 2. Shared communities between topics (and their word contexts) from El Observador relative to proposal for governance documents from Partido Nacional and Frente Amplio. El Observador shared more communities with Partido Nacional relative to Frente Amplio. This table summarizes results shown on Figure 11.
From Figure 11 and Table 2 the author found that topics (and their associated word contexts) discussed by El Observador newspaper regarding the political debates between Luis Lacalle Pou and Daniel Martínez shared more communities with the proposal for governance document from Partido Nacional relative to Frente Amplio. This would indicate an ideological bias towards Partido Nacional from the journalistic reporting of the debate by El Observador newspaper. Furthermore, topics in common that were shared by El Observador and Partido Nacional within the same communities included 'trabajo', 'gobernar', 'democracia' and 'mujer'.
Thus, the author has arrived at a methodology to measure the ideological leaning of a newspaper based on the word context used around topic terms and its similarity to word context from topic terms found at documents of governance from political parties. This methodology applied to the other four newspapers would serve to identify their political leaning and thus evaluate ideological diversity of the printed media in Uruguay.
In Figure 12 and Table 3 are shown the word contexts for topic terms derived from Montevideo Portal news articles written on the political debates.
Figure 12. Network graph visualization of the relationship between topics (and their word contexts) for Montevideo Portal newspaper and documents for governance from Partido Nacional and Frente Amplio, respectively. About 1,786 nodes and 2,965 edges were visualized using Gephi . Topics were arranged into 12 communities, and each community was color coded. Notice the relationship between nodes (corresponding topics) from Montevideo Portal (MPDeb) relative to nodes associated with documents for governance from Partido Nacional (PNprog) and Frente Amplio (FAProg), respectively. News articles written about both political debates (Oct-1 and Nov-13) between Luis Lacalle Pou and Daniel Martínez were used to extract the word contexts associated to 14 topics. The relationship between topics and community structure in the graph is shown more clearly on Table 3 below.
Table 3. Shared communities between topics (and their word contexts) from El Montevideo Portal relative to proposal for governance documents from Partido Nacional and Frente Amplio. Montevideo Portal shared more communities with Partido Nacional than with Frente Amplio. This table summarizes results shown on Figure 12.
From Figure 12 and Table 3 it can be seen that Montevideo Portal shared more communities of topic terms with document of governance from Partido Nacional rather than with Frente Amplio. This result suggested ideological bias of Montevideo Portal towards Partido Nacional. Topics that were shared within the same community among Montevideo Portal and both documents of governance were 'mujeres' and 'educación'. Interestingly, debate news articles from Montevideo Portal did not contain topics such as 'mujer' and 'reforma' and thus were not included in the network graph visualization shown on Figure 12.
Figure 13. Network graph visualization of the relationship between topics (and their word contexts) for La República newspaper and documents for governance from Partido Nacional and Frente Amplio, respectively. About 1,881 nodes and 3,172 edges were visualized using Gephi . Topics were arranged into 12 communities, and each community was color coded. Notice the relationship between nodes (corresponding topics) from La República (RepDeb) relative to nodes associated with documents for governance from Partido Nacional (PNprog) and Frente Amplio (FAProg), respectively. News articles written about both political debates (Oct-1 and Nov-13) between Luis Lacalle Pou and Daniel Martínez were used to extract the word contexts associated to 14 topics. The relationship between topics and community structure in the graph is shown more clearly on Table 4 below.
Table 4. Shared communities between topics (and their word contexts) from La República relative to proposal for governance documents from Partido Nacional and Frente Amplio. La República shared more communities with Frente Amplio relative to Partido Nacional. This table summarizes results shown on Figure 13.
From data shown on Figure 13 and Table 4 the author concluded that La República displayed ideological bias towards Frente Amplio. The topic shared within the same community between Frente Amplio document of governance and La República was 'gobernar'. Interestingly, there were topic terms missing from La República news articles on the debate that were present on news articles written during the primary elections. These terms were 'democracia', 'mujer' and 'mujeres', respectively and thus were not included on the network graph visualization shown on Figure 13.
Figure 14. Network graph visualization of the relationship between topics (and their word contexts) for La Diaria newspaper and documents for governance from Partido Nacional and Frente Amplio, respectively. About 1,879 nodes and 3,188 edges were visualized using Gephi . Topics were arranged into 12 communities, and each community was color coded. Notice the relationship between nodes (corresponding topics) from La Diaria (DiaDeb) relative to nodes associated with documents for governance from Partido Nacional (PNprog) and Frente Amplio (FAProg), respectively. News articles written about both political debates (Oct-1 and Nov-13) between Luis Lacalle Pou and Daniel Martínez were used to extract the word contexts associated to 14 topics. The relationship between topics and community structure in the graph is shown more clearly on Table 5 below.
Table 5. Shared communities between topics (and their word contexts) from La Diaria relative to proposal for governance documents from Partido Nacional and Frente Amplio. La Diaria shared equal number of communities with Partido Nacional and Frente Amplio. This table summarizes results shown on Figure 14.
From data shown on Figure 14 and Table 5 the author concluded that La Diaria displayed a balanced ideology when reporting on political debates. This could be explained -in part- because La Diaria published several news articles from the organization Verificado.uy (https://verificado.uy) that works to diminish disinformation campaigns by correcting outdating data mentioned by the politicians during the debate and publishing the corresponding corrections. The topic terms missing from La Diaria news articles on the debate that were present on news articles written during the primary elections were 'democracia' and 'gobernar'. These topics terms were not included on the network graph visualization shown on Figure 13.
Figure 15. Network graph visualization of the relationship between topics (and their word contexts) for La Red21 newspaper and documents for governance from Partido Nacional and Frente Amplio, respectively. About 1,879 nodes and 3,188 edges were visualized using Gephi . Topics were arranged into 12 communities, and each community was color coded. Notice the relationship between nodes (corresponding topics) from La Red21 (Red21Deb) relative to nodes associated with documents for governance from Partido Nacional (PNprog) and Frente Amplio (FAProg), respectively. News articles written about both political debates (Oct-1 and Nov-13) between Luis Lacalle Pou and Daniel Martínez were used to extract the word contexts associated to 14 topics. The relationship between topics and community structure in the graph is shown more clearly on Table 6 below.
Table 6. Shared communities between topics (and their word contexts) from La Red21 relative to proposal for governance documents from Partido Nacional and Frente Amplio. La Red21 shared more number of communities with Frente Amplio than with Partido Nacional. This table summarizes results shown on Figure 15.
From Figure 15 and Table 6 it can be seen that La Red21 shared more communities with Frente Amplio compared with Partido Nacional and thus displayed ideological bias towards Frente Amplio. Topics terms such as 'reforma', 'mujer' and 'democracia' were absent from La Red21 newspapers and thus they were not included in Figure 15 and Table 6. The topic term 'seguridad' (and its associated word context) shared the same community for the three documents.
The author concluded then that the approach above explained was useful in addressing the ideological bias for the journalistic reporting of political debates between Luis Lacalle Pou (Partido Nacional) and Daniel Martínez (Frente Amplio) for the five newspapers under consideration (Figure 16).
Figure 16. Detection of ideological bias in newspapers from Uruguay. Bias was measured relative to similarity in word context from topic terms in news articles written about political debates relative to documents for governance by Frente Amplio (socialist, progressive) and Partido Nacional (conservative), respectively.
In this work the author described the use of natural language processing and unsupervised machine learning algorithms for the analysis of news articles published by five newspapers from Uruguay concerning to their reporting of two political candidates (Luis Lacalle Pou and Daniel Martínez) running for president in the 2019 election cycle. The period of study addressed the journalistic reporting prior (June 1 to June 29), during (June 30) and after the primaries (July 1 to July 25); as well as both political debates between the opposing candidates held on October 1 (prior to the National Election on October 27) and November 13 (prior to the Balotaje on November 24), respectively.
The author's interest in analyzing how the media portrayed both candidates during their election campaigns was based on the possibility for detection of ideological bias during reporting by newspapers. Because of findings from previous studies describing that media (TV an radio) ownership in Uruguay was concentrated in a small group of private corporations [12, 13, 16] that may be jeopardizing democracy ought to low levels of ideological diversity as well as lack of plurality in communicating public agendas and political activities, the author thought important to address the ideological diversity of the printed press (newspapers) during critical moments of the election cycle. Although the political affiliation of the printed press in Uruguay has previously been acknowledged [14, 15], there has not been a quantitative study and empirical demonstration of ideological bias in journalistic reporting until now.
With this aim the author embarked on the exhaustive analysis of news articles that led him to several findings. First, the predominance of Daniel Martínez and Frente Amplio as the candidate of the respective incumbent political party was evidenced in the degree of cross-mentions when analyzing news articles searched and downloaded using Luis Lacalle Pou as keyword (Figure 4 and 5), relative to cross-mentions of Luis Lacalle Pou in news articles searched and downloaded using Daniel Martínez as keyword. There might be two reasons that could explain this finding; one is that the press may have framed their reporting in a context in which the ruling party was actively involved and thus mentioned when the opposing candidate and his respective political party were written about. This would be in accord with previous findings describing preferential attention towards the ruling party by the printed press . A second reason lies in the predominant style of reporting in Uruguay which is 'declarative'  and would entail the citation of Frente Amplio and its politicians when reporting on Luis Lacalle Pou as its political discourse was centered in a fierce opposition and constant attack on the ruling government. Despite this fact, Luis Lacalle Pou was still the most mentioned politician in the primaries (Figure 3).
Mentions in the press for running candidates of each political party (Figure 1 and 2) reflected the outcome of the primaries; with Luis Lacalle Pou's nomination followed by Juan Sartori in the second place for Partido Nacional, and Daniel Martínez's nomination followed by Carolina Cosse in the second place for Frente Amplio. Interestingly, Montevideo Portal had a more balanced reporting on candidates of Frente Amplio during the primaries (Figure 2) as the number of mentions of Carolina Cosse, Mario Bergara and Oscar Andrade was notoriously higher compared to El Observador and La República newspapers. After the primaries on June 30 of 2019, the election of running mate for the vice-presidency for each political party was reported by the press with varied intensity. For instance, the selection of Beatriz Argimón as running mate of Luis Lacalle Pow did not appear to generate much interest as evidenced by the number of mentions (Figure 1 and 3); that greatly contrasted with the selection of Graciel Villar as running mate of Daniel Martínez which amounted to many more mentions in the press (Figure 2 and 3).
The reporting style of La República during the primaries was puzzling and quite different from the other two newspapers. It showed an almost complete lack of interest towards reporting on Daniel Martínez as the primary candidate of Frente Amplio (only 5 news articles could be retrieved) and focused instead on reporting on Luis Lacalle Pou. This was not the case when reporting on both political debates (Figure 6). This misbalanced reporting from La República prompted the author to assemble a new dataset (Table 1) containing news articles only on both political debates. For this task, the author also included news articles from two other newspapers. Nonetheless, the dataset of news articles from the primaries was quite useful for the application of clustering and topic modeling algorithms (Figure 7 and 8) in order to identify the topics discussed during the primaries. The topics were later used to analyze their word contexts in news articles written on the political debates as means to asses the ideological bias towards one candidate or the other by the media sources in question.
The analysis of word contexts is much more powerful (and meaningful) than the analysis of single words and their frequencies because meaning and concepts are expressed as group of words in discrete sequences we know as sentences. For instance, extracting commonly co-occurring words from text (bigrams) can discover interesting patterns that emerge when news articles from one newspaper is compared to others . Similarly, documents containing proposal for governance from each political party can reveal differences in word usage (and thus intentions) that can be reflected in the word usage by newspapers (Figure 17). Two words that appeared together quite often across several documents during the primaries were 'fake news'; an important phenomenon that affected the political campaigns as misinformation was reported to be a new phenomenon in this election cycle.
Figure 17. Co-occurrances of words from news articles (primaries and debates) published by five newspapers and documents of governance for Partido Nacional and Frente Amplio.
The author performed document similarity through clustering in order to asses the topics discussed by newspapers during the primaries. A set of common topics were identified from which 14 topics were selected. These topics and their word contexts were then compared between documents of governance from each political party to news articles published by five newspapers on the political debate. New articles published on the political debated offered an invaluable opportunity to compare word usage from journalistic output to word usage written by political parties in their governance documents. Indeed, differences in frequency of mentions for these 14 topics in documents of governance suggested distinct interests from each political party (Figure 18). For instance, the topic 'seguridad' was the only one whose relative frequency was higher in document of governance from Partido Nacional in comparison to Frente Amplio.
Figure 18. Percentage of mentions relative to document length (in tokens) for each of 14 topics identified through topic modeling from news articles published on the primaries by El Observador, Montevideo Portal, and La República.
The word contexts for the topics shown on Figure 18 were investigated through concordance analysis for each document of governance and news articles published on the political debate by the five newspapers shown on Figure 6. The word context for each topic was visualized into a network graph (Figure 11 to 15) and the community structure for each topic was analyzed (Table 2 to 6). Sharing of words derived from the context of topics from different sources (document of governance and news articles for instance) would place the topics within the same community. Thus, counting the number of communities shared between newspapers with documents of governance from Partido Nacional and Frente Amplio allowed the author to establish similarities in word usage between the newspaper and the political party in question. This in fact could be a measure of political leaning (and ideological bias) of the newspaper and its reporting on the debate towards one political party or the other. In this manner the author could establish that La República and La Red21 displayed ideological bias towards Frente Amplio, whereas Montevideo Portal and El Observador displayed ideological bias towards Partido Nacional (Figure 16). La Diaria was ideologically balanced ought probably in part to the publication of 'data correction' articles from Verificado.uy.
This is the first report that empirically addressed ideological bias from the printed press in Uruguay.
 Krings, M (2018). Study shows how media's influence on public opinion varies by country. Kansas University News Service, Accessed on September 26 of 2019, URL: https://news.ku.edu/2018/09/13/study-shows-how-media-influence-public-individual-opinions-varies-country-factors
 Vu, HT; Jiang, L; Cueva Chacón, LM; Riedl, MJ; Tran, DC; Bobkowski, PS (2018). What influences media effects on public perception? A cross-national study of comparative agenda setting. International Communication Gazette, Accessed on September 26 of 2019, https://doi.org/10.1177/1748048518817652
 Happer, C; Philo, G (2013). The role of the media in the construction of public belief and social change. Journal of Social and Political Psychology, Vol. 1 (1): 321-336
 King, G; Schneer, B; White, A (2017). How the news media activate public expression and influence national agendas. Science 358: 776-780
 Gentzkow, M and Shapiro, JM (2010). What drives media slant? Evidence from U.S daily newspapers. Econometrica, 78(1): 35-71
 Niven, D (2003). Objective evidence on media bias: newspaper coverage of congressional party switchers. J&MC Quarterly, 80(2): 311-326
 Hamborg, F; Donnay, K; Gipp, B (2018). Automated identification of media bias in news articles: an interdisciplinary literature review. International Journal on Digital Libraries, https://doi.org/10.1007/s00799-018-0261-y
 Lazaridou, K; Krestel, P (2016). Identifying political bias in news articles. Bulleting of the IEEE TCDL 12
 https://en.wikipedia.org/wiki/Ideology - Accessed on October 1 of 2019
 Cooper, M (2002). Media ownership and democracy in the digital information age. Center for Internet & Society, Standford Law School, https://cyberlaw.stanford.edu/attachments/mediabooke.pdf - Accessed on October 1 of 2019
 Armijo, E (2009). Media ownership regulation: a comparative perspective. Georgia Journal of International and Comparative Law, Volume 37 (3): 421-468
 Nalvarte, P (2017). Investigación sobre concentración de medios en Uruguay señala numerosos casos que deben adaptarse a la ley de comunicación. Knight Center for Journalism in the Americas, https://knightcenter.utexas.edu/es/blog/00-19082-investigacion-sobre-concentracion-de-medios-en-uruguay-senala-numerosos-casos-que-debe - Accessed on October 1 of 2019
 Barrios, L (2019). La concentración de medios en Uruguay amenaza a la democracia. Caras & Caretas, https://www.carasycaretas.com.uy/concentracion-de-medios-amenaza-democracia/ - Accessed on October 2 of 2019
 Garcè, A (2012). De Marcha a Búsqueda: periodismo, partidos y política. Diario El Observador, 14 de Noviembre de 2012, https://www.elobservador.com.uy/nota/de-marcha-a-busqueda-periodismo-partidos-y-politica-201211149460 - Accessed on October 2 of 2019
 Desena, M (2016). La transiciòn democràtica en la prensa escrita uruguaya (1985-1989). http://www.geipar.udelar.edu.uy/wp-content/uploads/2016/09/Marcelo%20Desena_La%20transicion%20democratica%20en%20la%20prensa%20escrita%20uruguaya.pdf - Accessed on October 3 of 2019
 Schuliaquer, I (2017). El sistema mediàtico nacional de Uruguay. IX Congreso Latinoamericano de Ciencia Política, Montevideo, 26-28 de Julio
 https://country.eiu.com/article.aspx?articleid=137115797&Country=Uruguay&topic=Summary&subtopic=Political+structure - Accessed on October 14 of 2019
 Elfardy, H; Diab, M; Callison-Burch, C (2015). Ideological perspective detection using semantic features. Proceedings of the Fourth Joint Conference on Lexical and Computational Semantics (*SEM 2015), June 4-5: 137-146
 Kulkarni, V; Ye, J; Skiena, S; Yang Wang, W (2018). Multi-view models for political ideology detection of news articles. arXiv: 1809.03485v1
 Iyyer, M; Enns, P; Boyd-Graber, J; Resnik, P (2014). Political ideology detection using recurrent neural networks. Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, June 23-25: 1113-1122
 Ahmed, A; Xing, EP (2010). Staying informed: supervised and semi-supervised multi-view topical analysis of ideological perspective. Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, October 9-11: 1140-1150
 Xie, P; Xing, EP (2013). Integrating document clustering and topic modeling. arXiv: 1309.6874
 Bengfort, B; Bilbro, R; Ojeda, T (2018). Applied text analysis with Python. Published by O'Reilly Media Inc., Sebastopol, California, USA
 Subedi, S (2019). Text clustering with KMeans algorithm using Scikit-Learn. https://sanjayasubedi.com.np/nlp/nlp-with-python-document-clustering/ - Accessed on October 21 of 2019
 Subedi, S (2019). Topic modeling in Python using Scikit-Learn. https://sanjayasubedi.com.np/nlp/nlp-with-python-topic-modeling/ - Accessed on October 21 of 2019
 Russell, MA; Klassen, M (2019). Mining the social web. Published by O'Reilly Media Inc., Sebastopol, California, USA