2021-02-25: Computation + Journalism 2021 Trip Report

By -

One year ago, I was eagerly anticipating a trip to Boston in mid-March to attend the Computation + Journalism (C+J) Symposium 2020 hosted by Northeastern University.  Our PhD student Alexander Nwala would be presenting a poster on his Storygraph work, and I was especially looking forward to the keynote addresses, in particular one from Amanda Cox, whose work I've highlighted in my Information Visualization courses for several years. Unfortunately, this was the first of several conferences for me to be impacted by the coronavirus pandemic.  So, I was thrilled that the organizers decided to revive the 2020 program as a virtual symposium this year (program and schedule). Videos are promised, and I'll update this post when that's available.  [5/7/21: Recordings of the keynote speakers and panels are now available!]

Several of our WS-DL group members attended C+J 2021 on Friday, February 19, 2021 (thanks for the free registration!). The symposium was hosted on the ohyay.co platform, which provided an interesting interaction experience. One of the most noticeable features was that virtual applause was audible, which was remarked on by several speakers.

One side-effect of the symposium being postponed for almost a whole year (and that year being fairly eventful) was that the program of papers and contributed talks was heavily related to COVID and the 2020 US Presidential Election, but those major stories did shape much of journalism in 2020.

Besides the keynotes and invited panels, the remaining content was held in parallel sessions.  The biggest problem I had during the entire symposium was in deciding which sessions to attend; they all contained interesting work.  I didn't take comprehensive notes, but I'll briefly link to the work from some of the sessions that I was able to catch.  And I'll include some screenshots and tweets related to what I saw. In addition to these, there was good activity on Twitter at the #CJ2021 and #cplusj21 hashtags.

Keynote I

The program kicked off with a keynote from Amanda Cox, editor of The New York Times's The Upshot section and former graphics editor at the NYT.  Her talk focused on uncertainty in data and on the challenges in presenting uncertainty in data via visualizations to readers of The New York Times.

Session I

Rob Wells, Katy Seiter, Mary Hennigan, Arkansascovid: How We Ran A COVID News Website Through a College Journalism Class - ArkansasCovid.com (@arkansascovid) 

This was impressive work developed by a college journalism class.  The two student co-authors are now Assistant Editors for Arkansascovid.com.  See https://arkansascovid.com/about-2/ for more information on how this came about.

Derek Kravitz (@derekkravitz), Georgia Gee, Kyra Senese, Documenting COVID-19, collecting unstructured and confidential data from FOIA - https://documentingcovid19.io

Anyone can use the Documenting COVID-19 website to search state documents related to the pandemic obtained through open records laws and FOIA requests.

Alexander Nwala (@acnwala), Michele C. Weigle (@weiglemc), Michael L. Nelson (@phonedude_mln), 365 Dots in 2018, 2019, and 2020: Quantifying Attention of News Sources (slides, @storygraphbot)

Our work on Storygraph was presented both in Session I and also during the poster session.  We were grateful for the multiple opportunities to talk about this work with attendees.

Keynote II

The second keynote was from Deen Freelon (@dfreelon), Associate Professor in the UNC Hussman School of Media and Journalism. I liked the idea of these personalized filter maps to assess the state of your own filter bubble.  The chart below plots ideology from left to right on the x-axis (and also on a blue-red color scale), "truth score" on the y-axis (higher is better), and sizes the dots by frequency.

Posters

I missed the Session II presentations because I spent my time chatting with folks in the poster session.

Nick Hagar, Jack Bandy, Daniel Trielli, Yixue Wang, Nicholas Diakopoulos, Defining Local News: A Computational Approach

This work looked at classifying news outlets as local, regional, or national based on the outlet's follow count by distance from the origin of the outlet. This could have applications for detecting "pink slime" local news outlets.  This work also reminded me of Alexander Nwala's work on the Local Memory Project and its associated dataset of US and international local news outlets.

Priyanka Nanayakkara and Jessica Hullman, Toward Better Communication of Uncertainty in Science Journalism
This study looked at how uncertainty is talked about in science news writing and categorized particular uncertainty terms to compare how they're used in "great" writing versus "typical" writing. Priyanka and I talked about how over the past year, the general public has gotten a bit more used to hearing about uncertainty in science writing, whether they react appropriately to that uncertainty or not. 

Shun Yamaya, Saumya Bhadani, Filippo Menczer, Brendan Nyhan, Giovanni Luca Ciampaglia, Political audience diversity and news quality
This work investigated using audience diversity as a proxy to determine the quality of a news source. They found that having a diverse audience in terms of partisanship pointed to the news source having higher journalistic standards.

Jennifer Stromer-Galley, Brian McKernan, Sarah Bolden, Jeff Hemsley, Communicating the Facebook Political Ad Library: Data and Design Challenges - https://illuminating.ischool.syr.edu/campaign_2020/
This work provides a front-end to exploring Facebook's Political Ad Library. It includes tons of filters and facets for exploring the dataset, focusing on the 2020 Presidential Election. I had a great conversation with Jenny about the role of web archiving in the ability of researchers to study online advertisements in the future.

Invited Panel I

David Byler (@databyler), Micah Cohen (@micahcohen), Natalie Jackson (@nataliemj10), Nick Diakopoulos (@ndiakopoulos), Political Forecasting Meets Journalism 

This panel of journalists and researchers from The Washington Post, FiveThirtyEight, PRRI (Public Religion Research Institute), and Northwestern provided great insight into how journalists interact with and interpret political forecasting.

Session III

 Dimitar Nikolov, Alessandro Flammini, Filippo Menczer, Right and left, partisanship predicts (asymmetric) vulnerability to misinformation (@OSoMe_IU)   

This work investigated if personal partisanship predicted a user's likelihood to share misinformation. They asked "are highly partisan users more likely to share misinformation?" and "are users in echo chambers more likely to share misinformation?".  Indeed, they found that degree of partisanship can predict vulnerability to misinformation.


Julia Angwin (@JuliaAngwin) and Surya Mattu (@suryamattu), The Citizen Browser Project — Auditing the Algorithms of Disinformation - https://themarkup.org/series/citizen-browser
Journalists from The Markup presented their work on the Citizen Browser Project, a series of investigative articles looking at how disinformation travels over social media. This was made possible through a study where users installed a custom web browser that allowed them to easily share social media posts they saw with The Markup.

Paper Panel III

Leonard Bronner (@lennybronner), Al Johri (@aljohri) and Jeremy Bowers (@jeremybowers). Predicting Elections using Live Data at ​The Washington Post

Data scientists and engineers from The Washington Post presented their work on forecasting voter turnout during an election. Before being used in the 2020 general election, it was tested on the 2019 Virginia House of Delegates and state Senate elections.

Keynote III

The final keynote was delivered by David Rothschild (@DavMicRot), an economist at Microsoft Research. His talk was "Polling, Misinformation, and the Mainstream Media" and focused on the 2020 Election.  I appreciated his cautions against "horserace coverage", the focus on which has been driving me crazy for the past few years. Another major point he made was that if you're worried about misinformation/disinformation, don't just focus on social media -- mainstream news plays an important role, too.

Session IV

Dylan Halpern (@dchalpern), Marynia Kolak, Xun Li, Qinyun Lin, Visualizing the Pandemic: 1 Year Later - https://theuscovidatlas.org

The group from The Center for Spatial Data Science at the University of Chicago talked about their work on the US COVID Atlas, presenting up-to-date and historical data on the COVID pandemic in the US, including state and county-based data cases, deaths, testing, and now vaccinations.

Invited Panel II

Jen Christiansen (@ChristiansenJen), Catherine D’Ignazio (@kanarinka), Jessica Hullman (@JessicaHullman), Alberto Cairo (@AlbertoCairo), Conveying a Clear Message and Uncertainty with Graphics 

The conference wrapped up with a terrific panel on conveying uncertainity in graphics.  I agree with the tweet below -- when the video comes out, make sure to watch this one!

This was a great conference with interesting talks that highlighted the impact that computation and visualization can have "in the real world".  I've always been interested in how people take in new information, and the combination of computer science and journalism is an exciting intersection for me. I'll definitely keep an eye out for this series in the future.
 
-Michele


Update 5/7/2021: Added link to recordings of the keynotes and panels (see first paragraph)

Comments

Post a Comment