Our current projects
CDP: Speaker Behaviors and Group Dynamics
Using the data made available from the Council Data Project we are currently researching municipal council discussion. Who speaks (Men, Women, Republican, Democrat, etc.)? How often? During which topics (housing, budgeting, etc.)? Under which council configurations and more (full council, subcommittee, etc.)? We additionally plan on looking into training a classifier to identify different types of interjections based off of audio and transcript data. Using interjection data can we extend Justice, Interrupted to the municipal level? Our dataset currently includes speaker-annotated, timestamped-sentences, transcripts of over 2300 meetings from across the US.
Eva Maxfield Brown, Nic Weber
[add links: Dataset, Our Method for Efficient Fine-Tuning of Speaker ID Models, Analysis in Progress]
CDP: Special Interest Groups
Using the data made available from the Council Data Project we are about to begin a project to understand the contributions of special interest groups in municipal meetings. This will involve using (and potentially training) named-entity-recognition models to identity political organizations and individuals which contribute in public comment or are explicitly thanked by council members. We further will examine the effectiveness of these organizations in contributing to the passage of legislation.
Eva Maxfield Brown, Nic Weber, Undergraduate and Masters Program Contributors
Chinese-language Internet Yellow Pages and web archives
How did the web start to speak Chinese? This project investigates how the World Wide Web was experienced and adopted by Chinese-speaking users across the globe through examining Chinese-language Internet directory books published around the late 90s and early 2000s. These directory books - often marketed as "yellow pages" for the Internet - provide website links sorted into hierarchical categories for users without substantial knowledge in English or computers. Today, while most URLs featured in the books no longer work, these directory books can nevertheless provide valuable clues for researchers to understand the adoption of the web by Chinese-speaking users across the globe. For the Autumn 2022 and Winter 2023 quarters, the goal of the research is to use the URL data extracted from the directory books to benchmark the archival rate of Chinese-language web pags in Internet Archive’s Wayback Machine.
Richard Lewei Huang
GigPlat
A longitudinal cohort study of platform-based labor, with the goal of understanding how formal employment classifications and working conditions impact attitudes towards formal or informal labor protections. Since 2019, the research team has conducted focus groups and interviews with independent contractors working for on-demand platforms in the Seattle area. The recent phases of this project focus on the experiences of ridehail and delivery drivers during the Covid-19 pandemic and evaluating the impact of emergency protections set by Seattle City Council. Pre-registration of research protocol on the Open Science Framework https://osf.io/w5z6u
Nic Weber (PI), Lindsey Schwartz
Soft search
In collaboration with other teams, we are looking at how many, and which, NSF funded projects produce research software and what that software may look like. Are many projects unsustainable? Are they extendable or reusable? How can we build a system to predict if a project proposal will produce software? All members of the PIT clinic are working together on this project.