Growing Our Data, Growing Our Network

Harmony Labs
Harmony Labs
Published in
4 min readNov 10, 2020

Complex questions continue to challenge our understanding of media systems and their influence on politics, society, and culture. What’s media’s role in shaping the deep cultural narratives that condition public understanding? How do these narratives constrain public capacity to address critical social issues, like poverty, the climate crisis, or immigration reform? How are narratives formed and how can they be tracked computationally?

As we pursued answers to these questions over the last few years, it became increasingly clear that existing approaches and infrastructure were not meeting the needs of the research community. In 2019, we started the Harmony Labs Research Network.

The core of this network is a community of research and practice and a unique data platform that includes reliable access to proprietary and hard-to-get data sets, state-of-the-art analytical capacities, a growing network of developer applications, data and subject matter expertise from our in-house data scientists, and industry-grade data infrastructure and technical support.

With recent expansions of our data offering to include comprehensive online news, TV, and Facebook ad datasets (and more), the number of researchers leveraging this corpus of data to investigate healthy media systems has grown significantly in 2020, too. Here’s what they’re up to.

14 new projects across academia, news, & non-profits

Analyzing networks of audience overlap across news sources
University of Pennsylvania

  • Project Leads: Sandra Gonzalez-Bailon (Associate Professor of Communications) and Tian Yang (Doctoral Candidate)
  • Department: Annenberg School of Communications
  • Dataset: TV and Web panel data

Project Ratio: Quantifying the production and consumption of news
University of Pennsylvania

  • Project Leads: Duncan Watts (Professor), Homa Hosseinmardi (Postdoctoral Research Fellow), David Rothschild (Economist)
  • Department: Annenberg School of Communications, Department of Computer and Information Science
  • Dataset: TV and Web panel data; TV closed captioning; Online News
  • Publication: Evaluating the fake news problem at the scale of the information ecosystem in Science Advances

Assessing the effectiveness of FB ad enforcement policies in Myanmar
Carter Center

Creating a Facebook political ad dashboard for accurate insight into digital ad spending
Center for Responsive Politics

  • Project Leads: Sarah Bryner (Director of Research and Strategy) and Jacob Hileman (Information Technology Director)
  • Department: Political Ads Research
  • Dataset: FB political ad data; using machine learning to add value to related content for use by network partners.

Finding common ground in groups with ideological differences: using language as a signal
University of Maryland

  • Project Leads: Philip Resnik (Professor)
  • Department: Department of Linguistics and the Institute for Advanced Computer Studies (UMIACS)
  • Dataset: Civic engagement data; Online news data

Identifying unethical election advertising and monitoring a platform’s response.
Global Witness

  • Project Leads: Sam Leon
  • Department: Digital threats to democracy campaign
  • Dataset: FB political ad data

Exploring advertiser transparency and microtargeting in online political ads
University of Southern California

Monitoring hateful and decisive discourse around elections
Institute for Strategic Dialogue (ISD)

Investigating the influence of ad delivery algorithms
Northeastern University

  • Project Leads: Alan Mislove (Professor of Computer Science) and Piotr Sapieżyński (Associate Research Scientist)
  • Department: Khoury College of Computer Science
  • Dataset: FB political ad data

Understanding the dynamics of information dissemination and consumption for COVID-19 and the 2020 election
Northeastern University

  • Project Leads: David Lazer (Professor), Damian Ruck (Postdoctoral Research Fellow), Briony Swire Thompson (Senior Research Scientist)
  • Department: Political Science, Khoury College of Computer Sciences, Network Science Institute
  • Dataset: Online news data

Building a FB political ad dashboard with crowd sourced data
New York University

Investigating political advertising tactics of the major candidates and third-party groups
Politico

Assessing the impact of local news deserts on consumption of political news
Stanford University

  • Project Leads: Greg Martin (Assistant Professor) and Shoshana Vasserman (Assistant Professor)
  • Department: Stanford Graduate School of Business
  • Dataset: Civic engagement data; Online news data

A tool for customizable, daily updates on new political advertisements
Who Targets Me

What’s Next

The work underway by these groups has the potential to unlock important insights about the mechanics of media influence and help inform the design of healthier media systems in the future. We’re looking forward to working with each of these groups to ensure their findings have the widest possible impact, ranging from traditional academic outputs to new formats like blogs, websites, and the press. Watch this space for updates.

In the coming months we’re also hard at work sourcing additional data sets that are valuable to these types of research efforts and additional research partners to leverage them in their work. Contact us if that might be you.

--

--

Harmony Labs
Harmony Labs

Researching and reshaping our relationship with media