Find research datasets worth reusing

opencc-by-nc-4.0Dec 2022View details →

Navigating News Narratives: A Media Bias Analysis Dataset

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.Data description: 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).Data Format: The format of data is:<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> Annotation Scheme: The annotation scheme is based on Active learning, which is Manual Labeling --> Semi-Supervised Learning --> Human Verifications (iterative process)<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> List of datasets used : 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.We also utilize publicly available data from the following links. Our Attribution to others. MBIC (media bias):  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>  Hyperpartisan  news: 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> Toxic comment classification: 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>.Jigsaw Unintended Bias: 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>.Age Bias : 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>Multi-dimensional news Ukraine: 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> Social biases: 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>  Goal of this dataset :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.If you use this dataset, please cite us.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>

opencc-by-4.0Dec 2021View details →

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis during the COVID-19 pandemic (01.2020-06.2020)

Sources:  <ul> <li>Euractiv</li> <li>The Conversation</li> <li>Politico Europe </li> <li>IEEE Spectrum </li> <li>Techforge </li> <li>Fastcompany </li> <li>The Guardian (Tech) </li> <li>Arstechnica </li> <li>Reuters </li> <li>Gizmodo </li> <li>ZDNet </li> <li>The Register </li> <li>The Verge </li> <li>TechCrunch </li> </ul>   Methodology The sentiment analysis has been prepared using VADER*, an open-source lexicon and rule-based sentiment analysis tool. VADER is specifically designed for social media analysis, but can be also applied for other text sources. The sentiment lexicon was compiled using various sources (other sentiment data sets, Twitter etc.) and was validated by human input. The advantage of VADER is that the rule-based engine includes word-order sensitive relations and degree modifiers. As VADER is more robust in the case of shorter social media texts, the analysed articles have been divided into paragraphs. The analysis have been carried out for the social issues presented in the co-occurrence exercise. The process included the following main steps: <ul> <li>The 100 most frequently co-occurring terms are identified for every social issue (using the co-occurrence methodology)</li> <li>The articles containing the given social issue and co-occurring term are identified</li> <li>The identified articles are divided into paragraphs</li> <li>Social issue and co-occurring words are removed from the paragraph</li> <li>The VADER sentiment analysis is carried out for every identified and modified paragraph</li> <li>The average for the given word pair is calculated for the final result</li> </ul> Therefore, the procedure has been repeated for 100 words for all identified social issues. The sentiment analysis resulted in a compound score for every paragraph. The score is calculated from the sum of the valence scores of each word in the paragraph, and normalised between the values -1 (most extreme negative) and +1 (most extreme positive). Finally, the average is calculated from the paragraph results. Removal of terms is meant to exclude sentiment of the co-occurring word itself, because the word may be misleading, e.g. when some technologies or companies attempt to solve a negative issue. The neighbourhood's scores would be positive, but the negative term would bring the paragraph's score down. The analysed paragraphs are selected the following way: <ul> <li>The articles containing the given social issue are identified</li> <li>The paragraphs containing the social issue are selected for sentiment analysis</li> </ul> *Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text. Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.

opencc-by-4.0Mar 2019View details →

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis Sources: Above 140k articles (01.2016-03.2019): <ul> <li>Gigaom 0.5%</li> <li>Euractiv 0.9%</li> <li>The Conversation 1.3%</li> <li>Politico Europe 1.3%</li> <li>IEEE Spectrum 1.8%</li> <li>Techforge 4.3%</li> <li>Fastcompany 4.5%</li> <li>The Guardian (Tech) 9.2%</li> <li>Arstechnica 10.0%</li> <li>Reuters 11%</li> <li>Gizmodo 17.5%</li> <li>ZDNet 18.3%</li> <li>The Register 19.5%</li> </ul> Methodology The sentiment analysis has been prepared using VADER*, an open-source lexicon and rule-based sentiment analysis tool. VADER is specifically designed for social media analysis, but can be also applied for other text sources. The sentiment lexicon was compiled using various sources (other sentiment data sets, Twitter etc.) and was validated by human input. The advantage of VADER is that the rule-based engine includes word-order sensitive relations and degree modifiers. As VADER is more robust in the case of shorter social media texts, the analysed articles have been divided into paragraphs. The analysis have been carried out for the social issues presented in the co-occurrence exercise. The process included the following main steps: <ul> <li>The 100 most frequently co-occurring terms are identified for every social issue (using the co-occurrence methodology)</li> <li>The articles containing the given social issue and co-occurring term are identified</li> <li>The identified articles are divided into paragraphs</li> <li>Social issue and co-occurring words are removed from the paragraph</li> <li>The VADER sentiment analysis is carried out for every identified and modified paragraph</li> <li>The average for the given word pair is calculated for the final result</li> </ul> Therefore, the procedure has been repeated for 100 words for all identified social issues. The sentiment analysis resulted in a compound score for every paragraph. The score is calculated from the sum of the valence scores of each word in the paragraph, and normalised between the values -1 (most extreme negative) and +1 (most extreme positive). Finally, the average is calculated from the paragraph results. Removal of terms is meant to exclude sentiment of the co-occurring word itself, because the word may be misleading, e.g. when some technologies or companies attempt to solve a negative issue. The neighbourhood's scores would be positive, but the negative term would bring the paragraph's score down. The presented tables include the most extreme co-occurring terms for the analysed social issue. The examples are chosen from the list of words with 30 most positive and 30 most negative sentiment. The presented graphs show the evolution of sentiments for social issues. The analysed paragraphs are selected the following way: <ul> <li>The articles containing the given social issue are identified</li> <li>The paragraphs containing the social issue are selected for sentiment analysis</li> </ul> *Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text. Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.   Files sentiments_mod11.csv sentiment score based on chosen unigrams sentiments_mod22.csv sentiment score based on chosen bigrams sentiments_cooc_mod11.csv, sentiments_cooc_mod12.csv, sentiments_cooc_mod21.csv, sentiments_cooc_mod22.csv combinations of co-occurrences: unigrams-unigrams, unigrams-bigrams, bigrams-unigrams, bigrams-bigrams

opencc-by-4.0Sep 2019View details →

Fed4Fire/CDN-X-ALL network metrics dataset for time series analysis in Media content delivery for 4G/5G networks

The following dataset was generated at VICOMTECH (https://www.vicomtech.org) under project/experiment CDN-X-ALL: "CDN edge-cloud computing for efficient cache and reliable streaming aCROSS Aggregated unicast-multicast LinkS". Project funded by Fed4FIRE+ OC5 (<a href="https://www.fed4fire.eu/">https://www.fed4fire.eu</a>) under grant 732638. The dataset provides network metrics captures across several days employing a GStreamer-based MPEG-DASH player running on an UE connected to a LTE network. Nitos LTE/OpenAirInterface (OAI) testbed (<a href="https://nitlab.inf.uth.gr/NITlab/nitos/lte">https://nitlab.inf.uth.gr/NITlab/nitos/lte</a>) was used to deploy the LTE network. CDN-like server/DASH Dataset -> Internet -> EPC/OAI -> eNodeB/OAI -> UE/DASH player The player downloads MPEG-DASH video files provided by Distributed DASH dataset (<a href="https://dash.itec.aau.at/distributed-dash-datset/">https://dash.itec.aau.at/distributed-dash-datset/</a>), a dataset for CDN-like experiments, and captures the following data: <ol> <li>Date: date when the data is collected</li> <li>Player: type of the player (in this case it is always "GStreamer")</li> <li>Num: identifier of the player</li> <li>URLVid: URL of the MPD file</li> <li>Latency: latency experienced by the player</li> <li>BW: bandwidth experienced by the player</li> <li>Quality: chosen DASH video representation</li> </ol> During the experiments, other players run in order to generate realistic media streaming traffic at the CDN-like servers. These players start playing by following Poisson or Pareto distribution. The dataset was used to train Machine Learning Time Series predictor in order to forecast network capabilities and can be used for further experimentation concerning time series analysis.

opencc-by-4.0Jul 2020View details →

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis (01.2016-12.2019)

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis Sources with weights: <ul> <li>Euractiv 5%</li> <li>The Conversation 5%</li> <li>Politico Europe 5 %</li> <li>IEEE Spectrum 5 %</li> <li>Techforge 5%</li> <li>Fastcompany 5%</li> <li>The Guardian (Tech) 12%</li> <li>Arstechnica 5%</li> <li>Reuters 5%</li> <li>Gizmodo 9%</li> <li>ZDNet 9%</li> <li>The Register 12%</li> <li>The Verge 9%</li> <li>TechCrunch 9%</li> </ul> Methodology The sentiment analysis has been prepared using VADER*, an open-source lexicon and rule-based sentiment analysis tool. VADER is specifically designed for social media analysis, but can be also applied for other text sources. The sentiment lexicon was compiled using various sources (other sentiment data sets, Twitter etc.) and was validated by human input. The advantage of VADER is that the rule-based engine includes word-order sensitive relations and degree modifiers. As VADER is more robust in the case of shorter social media texts, the analysed articles have been divided into paragraphs. The analysis have been carried out for the social issues presented in the co-occurrence exercise. The process included the following main steps: <ul> <li>The 100 most frequently co-occurring terms are identified for every social issue (using the co-occurrence methodology)</li> <li>The articles containing the given social issue and co-occurring term are identified</li> <li>The identified articles are divided into paragraphs</li> <li>Social issue and co-occurring words are removed from the paragraph</li> <li>The VADER sentiment analysis is carried out for every identified and modified paragraph</li> <li>The average for the given word pair is calculated for the final result</li> </ul> Therefore, the procedure has been repeated for 100 words for all identified social issues. The sentiment analysis resulted in a compound score for every paragraph. The score is calculated from the sum of the valence scores of each word in the paragraph, and normalised between the values -1 (most extreme negative) and +1 (most extreme positive). Finally, the average is calculated from the paragraph results. Removal of terms is meant to exclude sentiment of the co-occurring word itself, because the word may be misleading, e.g. when some technologies or companies attempt to solve a negative issue. The neighbourhood's scores would be positive, but the negative term would bring the paragraph's score down.   *Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text. Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.

opencc-by-4.0Dec 2021View details →

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis (01.2016-04.2021)

dryad40/100

Transient analysis of power loss density with time-harmonic electromagnetic waves in Debye media

Due to the complex permittivity, it is difficult to directly clarify the transient mechanism between electromagnetic waves and Debye media. To overcome above problem, the temporal relationship between the electromagnetic waves and permittivity is explicitly derived by applying the Fourier inversion and introducing the remnant displacement. With the help of the Poynting theorem and energy conservation equation, the transient power loss density is derived to describe the transient dissipation of electromagnetic field and the mechanism on phase displacement has been explicitly revealed. Besides, the unique solution can be obtained by applying the time-domain analysis method rather than involving the frequency-domain characteristics. The effectiveness of transient analysis is demonstrated by giving a comparison simulation on one-dimensional example.

opencc-zeroJan 2022View details →

opencc-by-4.0Mar 2022View details →

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

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. News outlets: <ul> <li>The Guardian</li> <li>Daily Mail</li> <li>Globo</li> <li>Estadao</li> </ul>

opencc-by-4.0Jul 2021View details →

Analysis of websites, social media pages and apps for the development of new strategies for increasing participation of women in clinical trials

For T2.5 of the i-CONSENT project, an analysis was undertaken of the strategies used to effectively communicate with women on the topic of women’s health or women’s health research, considering aspects such as tone, format and audience interaction. A total of 42 websites, social media pages and apps were included in the analysis. The attached document presents the findings from the data generation stage of this analysis.

dryad40/100

Transient analysis of power loss density with time-harmonic electromagnetic waves in Debye media

Open the record for dataset details and reuse information.

publicJan 2022View details →

opencc-by-4.0Aug 2014View details →

Newspaper review and analysis data on media coverage of budget issues in Nigeria

The data was collected from analysis of six newspaper publications to establish citizen engagement and media coverage of the budget discourse from 2009 to 2013 as part of the investigation of the use of the online national budget of Nigeria. The work was part of the 'Exploring  the  Emerging  Impacts  of  Open  Data  in  Developing  Countries'  (ODDC) research  project.

opencc-by-4.0Oct 2023View details →

Data from Nicolle et al. LC-HRMS study of Streptomyces sp. AgN23 Culture Media Extract. Study of AgN23 exometabolome and analysis of Arabidopsis metabolomic responses to the bacteria

This archive compiles several datasets related to studies of Streptomyces sp. AgN23 interaction with Arabidopsis thaliana. Ultra-high-performance liquid chromatography-high-resolution MS (UHPLC-HRMS) analyses were performed on a Q Exactive Plus quadrupole (Orbitrap) mass spectrometer, equipped with a heated electrospray probe (HESI II) coupled to a U-HPLC Ultimate 3000 RSLC system (Thermo Fisher Scientific, Hemel Hempstead, United Kigdom). For each biological sample, the RAW file obtained in ESI+ and ESI- mode were retrieved from the Xcalibur version 4.4 software and are deposited in separate sub-folders termed "RawPos" and "RawNeg". Each experimental cohort is grouped in a folder where the biological repeats can be retrieved, as well as QC (Quality Check, pool of all samples from the cohort), Blank samples and eventual alternative control such as Bennett, the mock control media of Streptomyces sp.  AgN23. The details regarding samples preparation, analytic parameters and mass spectrometry, statistical treatment and visualization of the data will be made available in the publication relating to this archive. The folder " AgN23-WT_AgN23-pSC004" contains chromatograms related to metabolomic study of Wild-type and pSC004-1, pSC004-10, pSC004-16 and pSC004-22 mutants of Streptomyces sp. AgN23. The folder " Col-0_AgN23" contains chromatograms related to metabolomic study of Arabidopsis thaliana Col-0 responses to colonization by Streptomyces sp. AgN23-WT. The folder " Col-0_pad3-1_AgN23" contains chromatograms related to metabolomic study of Arabidopsis thaliana Col-0 and the Arabidopsis pad3-1 mutant responses to colonization by Streptomyces sp. AgN23-WT. The folder " Col-0_pSC004" contains chromatograms related to metabolomic study of Arabidopsis thaliana Col-0 responses to colonization by Streptomyces sp. AgN23-WT and the AgN23 mutants pSC004-10 and pSC004-22. It should be noted that in the publication associated with this archive, the pSC004-1, pSC004-10, pSC004-16 and pSC004-22 mutants are referred to as ΔgbnB-1, ΔgbnB-2, ΔgbnB-3 and ΔgbnB-4, respectively.

opencc-by-4.0Sep 2024View details →

Dataset for "Geospatial analysis of toponyms in geotagged social media posts"

<h1>Geotagged Twitter posts dataset</h1> Dataset used for the research presented in the following paper: Takayuki Hiraoka, Takashi Kirimura, Naoya Fujiwara (2024) "Geospatial analysis of toponyms in geo-tagged social media posts". We collected georeferenced Twitter posts tagged to coordinates inside the bounding box of Japan between 2012-2018. The present dataset represents the spatial distributions of all geotagged posts as well as posts containing in the text each of 24 domestic toponyms, 12 common nouns, and 6 foreign toponyms. The code used to analyze the data is available on <a href="https://github.com/takayukihir/geotagged-tweets" target="_blank" rel="noopener">GitHub</a>. <h2>Data description</h2> <ul> <li><code>preprocessed_mcntlt7_selected/</code>: Number of geotagged twitter posts in each grid cell. Each csv file under this directory associates each grid cell (spanning 30 seconds of latitude and 45 secoonds of longitude, which is approximately a 1km x 1km square, specified by an 8 digit code <code>m3code</code>) with the number of geotagged tweets tagged to the coordinates inside that cell (<code>tweetcount</code>). <code>file_names.json</code> relates each of the toponyms studied in this work to the corresponding datafile (<code>all</code> denotes the full data). Note that these data files are modified from the v1.0.0 to exclude posts that contain seven or more mentions.</li> <li><code>population/population_center_2020.xlsx</code>: Center of population of each municipality based on the 2020 census. Derived from data published by the Statistics Bureau of Japan on <a href="https://www.stat.go.jp/data/kokusei/topics/topi135.html" target="_blank" rel="noopener">their website</a> (Japanese)</li> <li><code>population/census2015mesh3_totalpop_setai_area.csv</code>: Resident population in each grid cell based on the 2015 census. Derived from data published by the Statistics Bureau of Japan on <a href="https://www.e-stat.go.jp/gis/statmap-search?page=1&type=1&toukeiCode=00200521&toukeiYear=2015&aggregateUnit=S&serveyId=S002005112015&statsId=T000846" target="_blank" rel="noopener">e-stat</a> (Japanese)</li> <li><code>population/economiccensus2016mesh3_jigyosyo_jugyosya_area.csv</code>: Employed population in each grid cell based on the 2016 Economic Census. Derived from data published by the Statistics Bureau of Japan on <a href="https://www.e-stat.go.jp/gis/statmap-search?page=1&type=1&toukeiCode=00200553&toukeiYear=2016&aggregateUnit=S&serveyId=S002005112016&statsId=T000917" target="_blank" rel="noopener">e-stat</a> (Japanese)</li> <li><code>japan_MetropolitanEmploymentArea2015map/</code>: Shape file for the boundaries of Metropolitan Employment Areas (MEA) in Japan. See <a href="https://www.csis.u-tokyo.ac.jp/UEA/index_e.htm" target="_blank" rel="noopener">this website</a> for details of MEA.</li> <li><code>ward_shapefiles/</code>: Shape files for the boundaries of wards in large cities, published by the Statistics Bureau of Japan on <a href="https://www.e-stat.go.jp/gis/statmap-search?page=1&type=2&aggregateUnitForBoundary=A&toukeiCode=00200553&toukeiYear=2016&serveyId=A002005532016&coordsys=1&format=shape&datum=2011" target="_blank" rel="noopener">e-stat</a></li> </ul>

opencc-by-4.0Jul 2022View details →

Porous Media Modeling and Physical Property Analysis

This is a supplementary dataset for the submitted manuscript. The dataset includes CT samples, porous media models, permeability and elasticity result data, and Matlab code to implement the modeling.

opencc-by-4.0Jul 2021View details →

Data for manuscript: "Longitudinal Analysis of Sentiment and Emotion in News Media Headlines Using Automated Labelling with Transformer Language Models"

This data set contains automated sentiment and emotionality annotations of 23 million headlines from 47 popular news media outlets popular in the United States.  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.  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.  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 The Huffington Post or Breitbart, 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. 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. The list of compressed files in this data set is listed next: -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.  -models.rar contains the Transformer sentiment and emotion annotation models used in the analysis. Namely:  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 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  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 -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 -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/ -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

opencc-by-4.0May 2024View details →

Electric car subsidies in social media and parliament 2022–2023: analysis results

Electric car subsidies in social media and the Finnish parliament 2022–2023: comparative analysis results obtained using three methods for qualitative classification of textual data.

opencc-by-4.0May 2024View details →

Content analysis of Spanish Defense institutions' NATO eFP social media messages (2017-2022)

Content analysis of Spanish Defense institutions' social media messages about the NATO Enhanced Forward Forward (eFP) mission on Facebook, Twitter (X), Instagram and YouTube (2017-2022).

opencc-by-4.0Jun 2019View details →

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis (01.2016-04.2019)

Sentiment analysis of tech media articles using VADER package and co-occurrence analysis Sources with weights: <ul> <li>Euractiv 5%</li> <li>The Conversation 5%</li> <li>Politico Europe 5 %</li> <li>IEEE Spectrum 5 %</li> <li>Techforge 5%</li> <li>Fastcompany 5%</li> <li>The Guardian (Tech) 12%</li> <li>Arstechnica 5%</li> <li>Reuters 5%</li> <li>Gizmodo 9%</li> <li>ZDNet 9%</li> <li>The Register 12%</li> <li>The Verge 9%</li> <li>TechCrunch 9%</li> </ul> Methodology The sentiment analysis has been prepared using VADER*, an open-source lexicon and rule-based sentiment analysis tool. VADER is specifically designed for social media analysis, but can be also applied for other text sources. The sentiment lexicon was compiled using various sources (other sentiment data sets, Twitter etc.) and was validated by human input. The advantage of VADER is that the rule-based engine includes word-order sensitive relations and degree modifiers. As VADER is more robust in the case of shorter social media texts, the analysed articles have been divided into paragraphs. The analysis have been carried out for the social issues presented in the co-occurrence exercise. The process included the following main steps: <ul> <li>The 100 most frequently co-occurring terms are identified for every social issue (using the co-occurrence methodology)</li> <li>The articles containing the given social issue and co-occurring term are identified</li> <li>The identified articles are divided into paragraphs</li> <li>Social issue and co-occurring words are removed from the paragraph</li> <li>The VADER sentiment analysis is carried out for every identified and modified paragraph</li> <li>The average for the given word pair is calculated for the final result</li> </ul> Therefore, the procedure has been repeated for 100 words for all identified social issues. The sentiment analysis resulted in a compound score for every paragraph. The score is calculated from the sum of the valence scores of each word in the paragraph, and normalised between the values -1 (most extreme negative) and +1 (most extreme positive). Finally, the average is calculated from the paragraph results. Removal of terms is meant to exclude sentiment of the co-occurring word itself, because the word may be misleading, e.g. when some technologies or companies attempt to solve a negative issue. The neighbourhood's scores would be positive, but the negative term would bring the paragraph's score down. The presented tables include the most extreme co-occurring terms for the analysed social issue. The examples are chosen from the list of words with 30 most positive and 30 most negative sentiment. The presented graphs show the evolution of sentiments for social issues. The analysed paragraphs are selected the following way: <ul> <li>The articles containing the given social issue are identified</li> <li>The paragraphs containing the social issue are selected for sentiment analysis</li> </ul> *Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text. Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.

opencc-by-4.0Oct 2024View details →

Indonesian Foreign Policy towards Iran: Shia Hate speech on X social media with LSTM and SVM Analysis

This table and figure are integral components of research on Shia hate speech on social media X, a critical issue in the context of identity politics in Indonesia and globally. This study is of paramount importance as it delves into the identity politics often exploited by politicians in Indonesia and around the world. The Shia community's support for President Jokowi in the first and second stages of the Election was met with hate speech from the opposition group. The study further investigates whether this Shia hate speech is linked to the government's policy towards Iran, a country known for its Shia ideology. The study is presented in three parts: 1. Table detailing the sentiment analysis process and results, which were conducted using advanced machine learning techniques such as SVM and LSTM. This approach significantly enhances the accuracy and reliability of the study's findings. 2. Figure in the form of a graph related to the study results and the results of comments from the Indonesian public about Shia. 3. This research is backed by a comprehensive dataset comprising public comments from Indonesia on Shia. This extensive data collection ensures the study's conclusions are thorough and reliable. 4. Processing of machine learning