Computational journalism has been an emerging discipline in recent years. A lot of inspiring data-driven studies on news production and consumption, news audience, news bias, and tools have been published and discussed. Some conferences, such as computation+journalism symposium, have started to offer places for computer scientists and journalists to help each other to discover new research directions.
On the other hands, social media has received a lot of attention from researchers who study public opinion. The data collected from social media is a valuable asset to see what people have in their mind about at the very moment. Election prediction based on Twitter is a good example of such efforts.
In this workshop, we are trying to make this two disciplines meet. How news media formulate public opinion and how public opinion influences on news media are our questions to ask. For this, of course, the parallel efforts on the understanding of news media and public opinion are required first, and then based on the findings, we can look into the interplay between them. As a result, the workshop topics can be categorized into three groups: news, public opinion, and their interplay.
Theme 1: Data-driven studies on news production and consumption. Theories on media influence have long been studied but in a limited setting mainly due to the lack of the data availability (previous studies heavily depends on survey results). We revisit some of the central questions in the media influence theories and aim to investigate them with a large-scale data. Relevant topics are (but not limited): News bias measurement, news coverage analysis, news diffusion analysis, news comment and content analysis, channel characteristics analysis, news story evolution tracking, automatic generation of news headline and text, news audience analysis, and news recommendation.
Theme 2: Accurate measurement of public opinion. Recently, web logs and/or social media have been used as a mean to capture the public opinion. Yet, it still suffers from the bias (e.g., selection bias, sampling bias, etc.). We open to new method to contribute for capturing public opinion, and we also seek to debiasing methods when sampling data from social media. The proposed methodology can also be useful for social sensing or nowcasting. Relevant topics are (but not limited): Measuring public opinion through social media, social sensing and nowcasting, Debiasing methods (e.g., re-weighting schemes) for social media data, summarizing and extracting sentiments, and connecting online and offline index.
Theme 3: Interplay between news and public opinion. By linking news and public opinion, we aim to quantify the impact of each of news article or news story and how it shapes public opinion. We will also discuss what is the practical implication of the research (e.g. for newsroom). Relevant topics are (but not limited): Impact of news on public opinion, influence of public opinion on news, and mapping between news and public opinion.
Special Theme 1: Fake news.
After the US presidential election, fake news has received a lot of attention from both academia and industry. There are many components regarding fake news: detecting fake news, users’ motivation of sharing fake news, the effect of fake news in shaping public opinion or the election result, filter bubbles and echo chambers of news feed in social media, and how to lessen the effect of misinformation.