Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
35
datasets available to search
ShareScore release 0.9.0
Dataset results
35 results for “news media”
Navigating News Narratives: A Media Bias Analysis Dataset
<p>The prevalence of bias in the news media has become a critical issue, affecting public perception on a range of important topics such as political views, health, insurance, resource distributions, religion, race, age, gender, occupation, and climate change. The media has a moral responsibility to ensure accurate information dissemination and to increase awareness about important issues and the potential risks associated with them. This highlights the need for a solution that can help mitigate against the spread of false or misleading information and restore public trust in the media.</p><p><strong>Data description: </strong>This is a dataset for news media bias covering different dimensions of the biases: political, hate speech, political, toxicity, sexism, ageism, gender identity, gender discrimination, race/ethnicity, climate change, occupation, spirituality, which makes it a unique contribution. The dataset used for this project does not contain any personally identifiable information (PII).</p><p><strong>Data Format: </strong>The format of data is:</p><ul><li>ID: Numeric unique identifier.</li><li>Text: Main content.</li><li>Dimension: Categorical descriptor of the text.</li><li>Biased_Words: List of words considered biased.</li><li>Aspect: Specific topic within the text.</li><li>Label: Neutral, Slightly Biased , Highly Biased</li></ul><p><br><strong>Annotation Scheme: </strong>The annotation scheme is based on Active learning, which is Manual Labeling --> Semi-Supervised Learning --> Human Verifications (iterative process)</p><ul><li>Bias Label: Indicate the presence/absence of bias (e.g., no bias, mild, strong).</li><li>Words/Phrases Level Biases: Identify specific biased words/phrases.</li><li>Subjective Bias (Aspect): Capture biases related to content aspects.</li></ul><p><br><strong>List of datasets used : </strong>We curated different news categories like Climate crisis news summaries , occupational, spiritual/faith/ general using RSS to capture different dimensions of the news media biases. The annotation is performed using active learning to label the sentence (either neural/ slightly biased/ highly biased) and to pick biased words from the news.</p><p>We also utilize publicly available data from the following links. Our Attribution to others.</p><p> <strong>MBIC (media bias): </strong>Spinde, Timo, Lada Rudnitckaia, Kanishka Sinha, Felix Hamborg, Bela Gipp, and Karsten Donnay. "MBIC--A Media Bias Annotation Dataset Including Annotator Characteristics." arXiv preprint arXiv:2105.11910 (2021). <a href="https://zenodo.org/records/4474336">https://zenodo.org/records/4474336</a> </p><p><strong>Hyperpartisan news: </strong>Kiesel, Johannes, Maria Mestre, Rishabh Shukla, Emmanuel Vincent, Payam Adineh, David Corney, Benno Stein, and Martin Potthast. "Semeval-2019 task 4: Hyperpartisan news detection." In Proceedings of the 13th International Workshop on Semantic Evaluation, pp. 829-839. 2019. <a href="https://huggingface.co/datasets/hyperpartisan_news_detection">https://huggingface.co/datasets/hyperpartisan_news_detection</a> </p><p><strong>Toxic comment classification: </strong>Adams, C.J., Jeffrey Sorensen, Julia Elliott, Lucas Dixon, Mark McDonald, Nithum, and Will Cukierski. 2017. "Toxic Comment Classification Challenge." Kaggle. <a href="https://kaggle.com/competitions/jigsaw-toxic-comment-classification-challenge">https://kaggle.com/competitions/jigsaw-toxic-comment-classification-challenge</a>.</p><p><strong>Jigsaw Unintended Bias: </strong>Adams, C.J., Daniel Borkan, Inversion, Jeffrey Sorensen, Lucas Dixon, Lucy Vasserman, and Nithum. 2019. "Jigsaw Unintended Bias in Toxicity Classification." Kaggle. <a href="https://kaggle.com/competitions/jigsaw-unintended-bias-in-toxicity-classification">https://kaggle.com/competitions/jigsaw-unintended-bias-in-toxicity-classification</a>.</p><p><strong>Age Bias : </strong>Díaz, Mark, Isaac Johnson, Amanda Lazar, Anne Marie Piper, and Darren Gergle. "Addressing age-related bias in sentiment analysis." In Proceedings of the 2018 chi conference on human factors in computing systems, pp. 1-14. 2018. <a href="https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/F6EMTS">Age Bias Training and Testing Data - Age Bias and Sentiment Analysis Dataverse (harvard.edu)</a></p><p><strong>Multi-dimensional news Ukraine: </strong>Färber, Michael, Victoria Burkard, Adam Jatowt, and Sora Lim. "A multidimensional dataset based on crowdsourcing for analyzing and detecting news bias." In Proceedings of the 29th ACM International Conference on Information & Knowledge Management, pp. 3007-3014. 2020. <a href="https://zenodo.org/records/3885351#.ZF0KoxHMLtV">https://zenodo.org/records/3885351#.ZF0KoxHMLtV</a> </p><p><strong>Social biases: </strong>Sap, Maarten, Saadia Gabriel, Lianhui Qin, Dan Jurafsky, Noah A. Smith, and Yejin Choi. "Social bias frames: Reasoning about social and power implications of language." arXiv preprint arXiv:1911.03891 (2019). <a href="https://maartensap.com/social-bias-frames/">https://maartensap.com/social-bias-frames/</a> </p><p> </p><p><strong>Goal of this dataset :</strong>We want to offer open and free access to dataset, ensuring a wide reach to researchers and AI practitioners across the world. The dataset should be user-friendly to use and uploading and accessing data should be straightforward, to facilitate usage.</p><p><strong>If you use this dataset, please cite us.</strong></p><p>Navigating News Narratives: A Media Bias Analysis Dataset © 2023 by <a href="https://www.linkedin.com/in/shainaraza/">Shaina Raza, Vector Institute </a>is licensed under <a href="http://creativecommons.org/licenses/by-nc/4.0/?ref=chooser-v1">CC BY-NC 4.0 </a></p><p> </p>
Dataset: Sentiment Analysis annotation of News headlines covering the Olympic legacy of Rio 2016 and London 2012 published by the Brazilian and British online media
<p>Dataset of 464 news headlines with sentiment manually annotated by a domain expert using the labels positive, negative and neutral. Data contains URLs for news articles published between 2004-2020 by the British and Brazilian media in English and Brazilian Portuguese covering the Olympic legacies of London 2012 and Rio 2016. Articles were collected from the news outlets’ websites using Google search engine.</p> <p>News outlets:</p> <ul> <li>The Guardian</li> <li>Daily Mail</li> <li>Globo</li> <li>Estadao</li> </ul>
Propaganda and fake news on the war in Ukraine: data from Russian-speaking social media communities
<p>The data set contains posts from social media networks popular among Russian-speaking communities. Information was searched based on pre-defined keywords ("war", "special military operation", etc.) and is mainly related to the ongoing war in Ukraine with Russia. After a thorough review and analysis of the data, both propaganda and fake news were identified. The collected data is anonymized. Feature engineering and text preprocessing can be applied to obtain new insights and knowledge from this data set. The data set is useful for the study of information wars and propaganda identification.</p>
News data for studying media exposure to the Boston Marathon bombings
<p>The news data sets released here have been used to study the relationship between media exposure and individuals' threat perception. Media exposure to mass violence has been shown to have a detrimental impact on people's threat perception and mental wellness, but little has been done to explore how exposure to different news content may impact mental health in people's everyday lives. In our study, we empirically test how emotionally potent media coverage of a real-world threat, namely, the Boston Marathon bombings occurred in 2013, alters threat perception of the community members over the first and the third anniversaries (in 2014 and 2016).</p> <p>The data were collected using a wave-based longitudinal design. There are two data sets, and each covers three waves:</p> <ul> <li>Dataset (I) -- news coverage before (Wave 1), during (Wave 2), and after (Wave 3) 2014 anniversary</li> <li>Dataset (II) -- news coverage before (Wave 1), during (Wave 2), and after (Wave 3) 2016 anniversary</li> </ul> <p>The collection procedure was informed by our survey study. Based on the survey completed by our subjects, we identified the four most frequent news outlets in the response: Metro (MT), New York Times (NY), Boston Globe (BG), and Boston Herald (BH). Other outlets, such as USA Today and Wall Street Journal, were reported by less than ten respondents. Therefore, our data collection focused on the news published by the four most frequent outlets.</p> <p>The data sets include the metadata of the news coverage over the aforementioned six waves. The raw content of the news stories was removed to respect the copyright owners.</p> <p><strong>Summary of the data collection procedure </strong></p> <p>We used news aggregators including Google and Yahoo news, to retrieve news articles published by the four outlets on a daily basis. We first collected the URLs of the news articles from the news aggregators and retrieved and parsed the news content using an HTML parser. In total, we collected over 38.5K and 54.1K news articles in dataset I and II, respectively.</p> <p>There are six files; each correspond to news coverage from the outlets in each wave. In these files, each line contains four columns: outlet, time, title, url which indicate the outlet of each news article, the time of publishing, the title of the article, and the URL to the article.</p> <p>We are making the data sets available for academic researchers and public use, to enable the discovery of new insights and development of better techniques to improve crisis communication and mental wellness.</p>
Reporting uncertified science in the news media during the Covid-19 pandemic
<p>Preprints have established a stable position in the dissemination of scientific findings. This position has been reinforced by the Covid-19 pandemic, which required the rapid dissemination of new scientific information. However, in most cases, preprints have not undergone peer review and, as a consequence, lack the scientific rigor of other scientific publications such as journal articles. This presents a challenge for journalists who are tasked with keeping the public informed about the latest scientific developments in the context of great uncertainty during a global pandemic. Having to report on rapidly changing circumstances under increasing pressure from social media while also having to compete for attention in a saturated media landscape might place strain on the adherence to journalistic norms. This does not only have implications in terms of the case-by-case accuracy of reporting, but on the public perception of science at large.</p> <p>This paper investigates the reporting of scientific information from pre-preprints based on a sample of 2,877 online news articles related to Covid-19 in the South African news media. Our results show that despite the publication of guidelines for reporting on preprints in the media, there is still a way to go regarding the judicious use of scientific information from preprints by journalists.</p> <p>The dataset includes the table of the 2,877 online news articles as well as the analysis done (as separate tables).</p> <p>Fields in the dataset: Full Text, Headline, ID, Publication Date, Site Name,URL,Syndicated, Covid-related scientific research findings,<br> Preprint, Title, Authors,DOI, Archive / journal, URL</p>
A dataset of media releases (Twitter, News and Comments, Youtube, Facebook) form Poland related to COVID-19 for open research
<p>Social behavior has a fundamental impact on the dynamics of infectious diseases (such as COVID-19), challenging public health mitigation strategies and possibly the political consensus. The widespread use of the traditional and social media on the Internet provides us with an invaluable source of information on societal dynamics during pandemics. With this dataset, we aim to understand mechanisms of COVID-19 epidemic-related social behavior in Poland deploying methods of computational social science and digital epidemiology. We have collected and analyzed COVID-19 perception on the Polish language Internet during 15.01-31.07(06.08) and labeled data quantitatively (Twitter, Youtube, Articles) and qualitatively (Facebook, Articles and Comments of Article) in the Internet by infomediological approach.</p> <p>- manually labelled1,449 articles / Facebook posts from Lower Silesia (facebook_articles_lower_silesia.zip) and 111 texts from outside this region;</p> <p>-manually labelled 1000 most popular tweets (twits_annotated.xlsx) with cathegories is_fake (categorical and numeric) topic and sentiment; </p> <p>-extracted 57,306 representative articles (articles_till_06_08.zip) in Polish using Eventregitry.org tool in language Polish and topic "Coronavirus" in article body;</p> <p>- extracted 1,015,199 (tweets_till_31_07_users.zip and tweets_till_31_07_text.zip) and Tweets from #Koronawirus in language Polish using Twitter API.</p> <p>- collected 1,574 videos (youtube_comments_till_31_07.zip and youtube_movie.csv) with keyword: Koronawirus on YouTube and 247,575 comments on them using Google API;</p> <p>- We supplemented the media observations with an analysis of 244 social empirical studies till 25.05 on COVID-19 in Poland (empirical_social_studies.csv).</p> <p>Reports and analyzes and coding books can be found in Polish at: <a href="http://www.infodemia-koronawirusa.pl/percepcja-koronowirusa-na-dolnym-slasku/">http://www.infodemia-koronawirusa.pl</a></p> <p>Main report (in Polish) https://depot.ceon.pl/handle/123456789/19215 </p>
NLP and machine learning to measure peace from news media
<p>"Hate speech" can mobilize violence and destruction. What are the characteristics of "peace speech" that reflect and support the social processes that maintain peace? In this study we used a data driven, machine learning approach to identify the words most associated with lower-peace versus higher-peace countries. Logistic regression and random forest classifiers were trained using five respected, traditional peace indices: Global Peace Index, Positive Peace Index, World Happiness Index, Fragile States Index, and Human Development Index. The feature inputs into the machine learning model were the word frequencies from the news media in each country and the output classifications were the level of peace in that country. The machine learning model was successful in properly classifying the level of peace from the news media in a country (both accuracy and F1: 96% - 100%). We also used that trained machine model to create a machine learning peace index that measured the level of peace in countries, including countries not in the training set, which correlated with the average of those five traditional peace indices (r-squared = 0.8349). Using the random forest feature importance method we found that the words in news media in lower-peace countries were characterized by words related to government, order, control and fear (such as government, state, law, security and court), while higher-peace countries were characterized by an increased prevalence of words related to optimism for the future and fun (such as time, like, home, believe and game).</p>
Conservative News Media and Criminal Justice: Evidence from Exposure to Fox News Channel
<p>Conservative News Media and Criminal Justice: Evidence from Exposure to Fox News Channel" (Elliott Ash and Michael Poyker), <i>Economic Journal</i>, 2023</p>
Data for manuscript "The Prevalence of Terms Denoting Far-right and Far-left Political Extremism in U.S. and U.K. News Media"
<p>This data set belongs to an academic manuscript examining longitudinally (2000-2019) the prevalence of terms denoting far-right and far-left political extremism in a large corpus of more than 32 million written news and opinion articles from 54 news media outlets popular in the United States and the United Kingdom.</p> <p>The textual content of news and opinion articles from the 54 outlets listed in the main manuscript is available in the outlet's online domains and/or public cache repositories such as Google cache (https://webcache.googleusercontent.com), The Internet Wayback Machine (https://archive.org/web/web.php), and Common Crawl (https://commoncrawl.org). We used derived word frequency counts from these sources. Textual content included in our analysis is circumscribed to articles headlines and main body of text of the articles and does not include other article elements such as figure captions.</p> <p>Targeted textual content was located in HTML raw data using outlet specific xpath expressions. Tokens were lowercased prior to estimating frequency counts. To prevent outlets with sparse text content for a year from distorting aggregate frequency counts, we only include outlet frequency counts from years for which there is at least 1 million words of article content from an outlet. This threshold was chosen to maximize inclusion in our analysis of outlets with sparse amounts of articles text per year. </p> <p>Yearly frequency usage of a target word in an outlet in any given year was estimated by dividing the total number of occurrences of the target word in all articles of a given year by the number of all words in all articles of that year. This method of estimating frequency accounts for variable volume of total article output over time.</p> <p>The list of compressed files in this data set is listed next:</p> <p>-analysisScripts.rar contains the analysis scripts used in the main manuscript </p> <p>-articlesContainingTargetWords.rar contains counts of target words in outlets articles as well as total counts of words in articles</p> <p> </p> <p>Usage Notes</p> <p>In a small percentage of articles, outlet specific XPath expressions failed to properly capture the content of the article due to the heterogeneity of HTML elements and CSS styling combinations with which articles text content is arranged in outlets online domains. As a result, the total and target word counts metrics for a small subset of articles are not precise. In a random sample of articles and outlets, manual estimation of target words counts overlapped with the automatically derived counts for over 90% of the articles.</p> <p>Most of the incorrect frequency counts were minor deviations from the actual counts such as for instance counting the word "Facebook" in an article footnote encouraging article readers to follow the journalist’s Facebook profile and that the XPath expression mistakenly included as the content of the article main text. Some additional outlet-specific inaccuracies that we could identify occurred in "The Hill" and "Newsmax" news outlets where XPath expressions had some shortfalls at precisely capturing articles’ content. For "The Hill", in years 2007-2009, XPath expressions failed to capture the complete text of the article in about 40% of the articles. This does not necessarily result in incorrect frequency counts for that outlet but in a sample of articles’ words that is about 40% smaller than the total population of articles words for those three years. In the case of "NewsMax", the issue was that for some articles, XPath expressions captured the entire text of the article twice. Notice that this does not result in incorrect frequency counts. If a word appears x times in an article with a total of y words, the same frequency count will still be derived when our scripts count the word 2x times in the version of the article with a total of 2y words.</p> <p>To conclude, in a data analysis of 32 million articles, we cannot manually check the correctness of frequency counts for every single article and hundred percent accuracy at capturing articles’ content is elusive due to the small number of difficult to detect boundary cases such as incorrect HTML markup syntax in online domains. Overall however, we are confident that our frequency metrics are representative of word prevalence in print news media content (see Figure 1 in the main manuscript for illustration of the accuracy of the frequency counts).</p>
A Social and News Media Benchmark Dataset for Topic Modeling
<p>A novel approach to topic modeling using PSO-based clustering is compared to traditional techniques using the 20 Newsgroups dataset and a collection of posts from the Reddit health forum r/Cancer.</p>
Dataset for Report: "The Increasing Prominence of Prejudice and Social Justice Rhetoric in UK News Media"
<p>This data set contains frequency counts of target words in 16 million news and opinion articles from 10 popular news media outlets in the United Kingdom. The target words are listed in the associated report and are mostly words that denote prejudice or are often associated with social justice discourse. A few additional words not denoting prejudice are also available since they are used in the report for illustration purposes of the method.</p> <p>The textual content of news and opinion articles from the outlets is available in the outlet's online domains and/or public cache repositories such as Google cache (https://webcache.googleusercontent.com), The Internet Wayback Machine (https://archive.org/web/web.php), and Common Crawl (https://commoncrawl.org). We used derived word frequency counts from these sources. Textual content included in our analysis is circumscribed to articles headlines and main body of text of the articles and does not include other article elements such as figure captions.</p> <p>Targeted textual content was located in HTML raw data using outlet specific xpath expressions. Tokens were lowercased prior to estimating frequency counts. To prevent outlets with sparse text content for a year from distorting aggregate frequency counts, we only include outlet frequency counts from years for which there is at least 1 million words of article content from an outlet. This threshold was chosen to maximize inclusion in our analysis of outlets with sparse amounts of articles text per year. </p> <p>Yearly frequency usage of a target word in an outlet in any given year was estimated by dividing the total number of occurrences of the target word in all articles of a given year by the number of all words in all articles of that year. This method of estimating frequency accounts for variable volume of total article output over time.</p> <p>In a small percentage of articles, outlet specific XPath expressions might fail to properly capture the content of the article due to the heterogeneity of HTML elements and CSS styling combinations with which articles text content is arranged in outlets online domains. As a result, the total and target word counts metrics for a small subset of articles are not precise. In a random sample of articles and outlets, manual estimation of target words counts overlapped with the automatically derived counts for over 90% of the articles.</p> <p>Most of the incorrect frequency counts are often minor deviations from the actual counts such as for instance counting the word "Facebook" in an article footnote encouraging article readers to follow the journalist’s Facebook profile and that the XPath expression mistakenly included as the content of the article main text.To conclude, in a data analysis of over 16 million articles, we cannot manually check the correctness of frequency counts for every single article and hundred percent accuracy at capturing articles’ content is elusive due to the small number of difficult to detect boundary cases such as incorrect HTML markup syntax in online domains. Overall however, we are confident that our frequency metrics are representative of word prevalence in print news media content (see Figure 2 of main manuscript for supporting evidence of the temporal precision of the method).</p>
Data for manuscript: "Longitudinal Analysis of Sentiment and Emotion in News Media Headlines Using Automated Labelling with Transformer Language Models"
<p>This data set contains automated sentiment and emotionality annotations of 23 million headlines from 47 popular news media outlets popular in the United States. </p> <p>The set of 47 news media outlets analysed (listed in Figure 1 of the main manuscript) was derived from the AllSides organization <a href="https://www.allsides.com/blog/updated-allsides-media-bias-chart-version-11">2019 Media Bias Chart v1.1</a>. The human ratings of outlets’ ideological leanings were also taken from this chart and are listed in Figure 2 of the main manuscript. </p> <p>News articles headlines from the set of outlets analyzed in the manuscript are available in the outlets’ online domains and/or public cache repositories such as The Internet Wayback Machine, Google cache and Common Crawl. Articles headlines were located in articles’ HTML raw data using outlet-specific XPath expressions. </p> <p>The temporal coverage of headlines across news outlets is not uniform. For some media organizations, news articles availability in online domains or Internet cache repositories becomes sparse for earlier years. Furthermore, some news outlets popular in 2019, such as <em>The Huffington Post</em> or <em>Breitbart</em>, did not exist in the early 2000’s. Hence, our data set is sparser in headlines sample size and representativeness for earlier years in the 2000-2019 timeline. Nevertheless, 18 outlets in our data set have chronologically continuous partial or full headline data availability fulfilling our inclusive criteria (see manuscript Methods) since the year 2000. Figure S 1 in the SI reports the number of headlines per outlet and per year in our analysis.</p> <p>In a small percentage of articles, outlet specific XPath expressions might fail to properly capture the content of the headline due to the heterogeneity of HTML elements and CSS styling combinations with which articles text content is arranged in outlets online domains. After manual testing, we determined that the percentage of headlines following in this category is very small. Additionally, our method might miss detecting some articles in the online domains of news outlets. To conclude, in a data analysis of over 23 million headlines, we cannot manually check the correctness of every single data instance and hundred percent accuracy at capturing headlines’ content is elusive due to the small number of difficult to detect boundary cases such as incorrect HTML markup syntax in online domains. Overall however, we are confident that our headlines set is representative of headlines in print news media content for the studied time period and outlets analyzed.</p> <p>The list of compressed files in this data set is listed next:</p> <p>-analysisScripts.rar contains the analysis scripts used in the main manuscript as well as aggregated data of sentiment and emotionality automated annotations of the headlines and human annotations of a subset of headlines sentiment and emotionality used as ground truth. </p> <p>-models.rar contains the Transformer sentiment and emotion annotation models used in the analysis. Namely: </p> <p>Siebert/sentiment-roberta-large-english from https://huggingface.co/siebert/sentiment-roberta-large-english. This model is a fine-tuned checkpoint of <a href="https://huggingface.co/roberta-large">RoBERTa-large</a> (<a href="https://arxiv.org/pdf/1907.11692.pdf">Liu et al. 2019</a>). It enables reliable binary sentiment analysis for various types of English-language text. For each instance, it predicts either positive (1) or negative (0) sentiment. The model was fine-tuned and evaluated on 15 data sets from diverse text sources to enhance generalization across different types of texts (reviews, tweets, etc.). See more information from the original authors at https://huggingface.co/siebert/sentiment-roberta-large-english</p> <p>DistilbertSST2.rar is the default sentiment classification model of the HuggingFace Transformer library https://huggingface.co/ This model is only used to replicate the results of the sentiment analysis with sentiment-roberta-large-english </p> <p>DistilRoberta j-hartmann/emotion-english-distilroberta-base from https://huggingface.co/j-hartmann/emotion-english-distilroberta-base. The model is a fine-tuned checkpoint of <a href="https://huggingface.co/distilroberta-base">DistilRoBERTa-base</a>. The model allows annotation of English text with Ekman's 6 basic emotions, plus a neutral class. The model was trained on 6 diverse datasets. Please refer to the original author at https://huggingface.co/j-hartmann/emotion-english-distilroberta-base for an overview of the data sets used for fine tuning. https://huggingface.co/j-hartmann/emotion-english-distilroberta-base</p> <p>-headlinesDataWithSentimentLabelsAnnotationsFromSentimentRobertaLargeModel.rar URLs of headlines analyzed and the sentiment annotations of the siebert/sentiment-roberta-large-english Transformer model. https://huggingface.co/siebert/sentiment-roberta-large-english</p> <p>-headlinesDataWithSentimentLabelsAnnotationsFromDistilbertSST2.rar URLs of headlines analyzed and the sentiment annotations of the default HuggingFace sentiment analysis model fine-tuned on the SST-2 dataset. https://huggingface.co/</p> <p>-headlinesDataWithEmotionLabelsAnnotationsFromDistilRoberta.rar URLs of headlines analyzed and the emotion categories annotations of the j-hartmann/emotion-english-distilroberta-base Transformer model. https://huggingface.co/j-hartmann/emotion-english-distilroberta-base</p>
Incentivizing news consumption on social media platforms using large language models and realistic bot accounts
<p>This project examines how to enhance users' exposure to and engagement with verified and ideologically balanced news in an ecologically valid setting. We rely on a large-scale two-week long field experiment on 28,457 Twitter users. We created 28 bots utilizing GPT-2 that replied to users tweeting about sports, entertainment, or lifestyle with a contextual reply containing two hardcoded elements: a URL to the topic-relevant section of quality news organization and an encouragement to follow its Twitter account. Treated users were randomly assigned to receive responses by bots presented as female or male. We examine whether our intervention enhances the following of news media organization, the sharing/liking of news content and the tweeting/liking of political content. We find that the treated users followed more news accounts and the users in the female bot treatment were more likely to like news content than the control.</p>
Data for manuscript "Prevalence in News Media of two Competing Hypotheses about COVID-19 Origins"
<p>The Covid-19 pandemic has been one of the most disruptive and painful phenomena of the last few decades. As of July 2021, the origins of the SARS-CoV-2 virus that caused the outbreak remain a mystery. This work analyzes the prevalence in news media articles of two popular hypotheses about SARS-CoV-2 virus origins: the natural emergence and the lab-leak hypotheses. </p> <p>This data set contains frequency counts of target words in news and opinion articles from 12 popular news media outlets. The target words are listed in the associated manuscript and are mostly words associated with the Covid-19 pandemic. </p> <p>The list of compressed files in this data set is listed next:</p> <p>targetWordsInArticlesCounts.rar contains counts of target words in outlets articles as well as total counts of words in articles</p> <p>targetWordsFrequencies.rar daily, weekly, monthly word frequencies</p> <p>wordEmbeddingModels.rar monthly embedding models of news outlets content</p> <p>analysisScripts.rar analysis notebooks</p> <p>The textual content of news and opinion articles from the outlets is available in the outlet's online domains and/or public cache repositories such as Google cache, The Internet Wayback Machine, and Common Crawl. We used derived word frequency counts from these sources. Textual content included in our analysis is circumscribed to articles headlines and main body of text of the articles and does not include other article elements such as figure captions.</p> <p>Targeted textual content was located in HTML raw data using outlet specific XPath expressions. Tokens were lowercased prior to estimating frequency counts. </p> <p>Yearly frequency usage of a target word in an outlet in any given temporal interval ( daily, weekly, monthly) was estimated by dividing the total number of occurrences of the target word in all articles of a given temporal interval by the number of all words in all articles of that temporal interval. This method of estimating frequency accounts for variable volume of total article output over time.</p> <p>In a small percentage of articles, outlet specific XPath expressions might fail to properly capture the content of the article due to the heterogeneity of HTML elements and CSS styling combinations with which articles text content is arranged in outlets online domains. As a result, the total and target word counts metrics for a small subset of articles are not precise. In a random sample of articles and outlets, manual estimation of target words counts overlapped with the automatically derived counts for over 90% of the articles. Most of the incorrect frequency counts are minor deviations from the actual counts such as for instance counting a word in an article footnote encouraging article readers to find related articles and that the XPath expression might mistakenly include as the content of the article main text. Some additional outlet-specific inaccuracies that we could identify occurred in the WSJ where in less than 5% of the articles XPath expressions failed to capture the article's main text content. Other outlets articles samples sizes might not be comprehensive but, to the best of our knowledge, they are representative and include tens of thousands of articles per outlet/year. To conclude, in a data analysis of over 1.5 million articles, we cannot manually check the correctness of frequency counts for every single article and hundred percent accuracy at capturing articles’ content is elusive due to the small number of difficult to detect boundary cases such as incorrect HTML markup syntax in online domains. Overall however, we are confident that our frequency metrics are representative of word prevalence in print news media content (see Figure 1 of main manuscript for supporting evidence).</p> <p> </p> <p> </p>
Changes in the use of 'refugees' and 'migrants' in international news media and academic publications 2010–2022
<p>The dataset contains source data for Figure 2 in the article, based on search results from Factiva and Web of Science.</p>
Network embedding for understanding the National Park System through the lenses of news media, scientific communication and biogeography
<p>The United States national parks encompass a variety of biophysical and historical resources important for national cultural heritage. Yet how these resources are socially constructed often depends upon the beholder. Parks tend to be conceptualized according to their (fixed) geographic context, so our understanding of this system of systems is dominated by this geographic lens. To expose the systemic structure that exists beyond their geographic embedding, we analyze three representations of the national park system using park-park similarity networks according to their co-occurrence in: (a) ~423,000 news media articles; (b) ~11,000 research publications; and (c) ~60,000 species inhabiting parks. We quantify structural variation between network representations by leveraging similarity measures at different scales: park-level (park-park correlations) and system-level (network communities' consistency). Because parks are governed and experienced at multiple scales, cross-network comparison informs how management should account for the varying objectives and constraints that dominate at each scale. Our results identify an interesting paradox: whereas park-level correlations depend strongly on the representative lens, the network communities are remarkably robust and consistent with the underlying geographic embedding. Our data-driven methodology is generalizable to other geographically embedded socio-environmental systems and supports the holistic analysis of systems-level structure that may elude other approaches.</p>
NLP and machine learning to measure peace from news media
Open the record for dataset details and reuse information.
Incentivizing news consumption on social media platforms using large language models and realistic bot accounts
Open the record for dataset details and reuse information.
Network embedding for understanding the National Park System through the lenses of news media, scientific communication and biogeography
Open the record for dataset details and reuse information.
Critical Gaps in Medical Research Reporting by Online News Media: Dataset
<p>Dataset for <strong>Critical Gaps in Medical Research Reporting by Online News Media</strong></p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.