As part of a research grant through the Chan Zuckerberg Initiative (2022-2024), we have developed Knowledge Map (KM), a reporting tool for survey data taken from past research on environmental challenges faced by communities in Dane County, Wisconsin. KM’s development was influenced by research with local government and non-profit organizations, which displays (1) the representativeness of the survey sample and (2) data visualizations. The representativeness is produced by comparing the survey responses’ distribution to a selected geographic level (e.g., census tract, county). It provides an evaluation of the representativeness of the data using the approach described in Qi et al. (2021). The visualizations report mostly categorical data with open-text responses using structured topic modeling. The tool has been developed using RShiny and is available in a private repository (see: ccs-knowledge-map); however, student turnover has left the project with several outstanding tasks for which we would benefit from DSI’s Open Source Internship program.
Internship Tasks:
Our project would benefit from tasks in three areas - coding/ feature development, data analysis, and documentation. However, not all are required for the internship.
Coding & Development: We have a list of backlog coding updates that need to be completed, ranging from minor improvements, e.g., fixing navigation, to feature development, e.g., integrating map visualizations. The log can be accessed here: KM Backlog (we are converting items from Notion). Key priorities include improving UI responsiveness and enhancing interactivity in visualizations.
Data Analysis: We want to improve our measures of representativeness. The current approach uses distributional comparisons across geographic units (e.g., census tracts, counties) to determine how well the survey sample reflects the broader population. However, this does not account for the sample size. We should improve this to warn users of low N or augment the model by incorporating uncertainty estimates or confidence intervals to indicate the reliability of representativeness scores. The approach can be found here:
See: Qi, Miao, Owen Cahan, Morgan A. Foreman, Daniel M. Gruen, Amar K. Das, and Kristin P. Bennett. “Quantifying representativeness in randomized clinical trials using machine learning fairness metrics.” JAMIA open 4, no. 3 (2021): ooab077.
Documentation & Open-Source Readiness: We must ensure Knowledge Map’s repository follows open-source best practices to make the project open-source. Our project would benefit from evaluating those practices: GitHub Repository and an overall assessment of code structure, licensing, contributor guidelines, and API documentation. Additionally, we should create onboarding materials for new developers and data users to encourage broader engagement with KM. Completing these would help enhance the accessibility and usability of the tool for researchers, policymakers, and community organizations, enabling them to make data-driven decisions with a clearer understanding of their data and its ability to represent all community members. This work would also strengthen the infrastructure of KM, ensuring it can be maintained and expanded upon in the future.
Mentorship:
The intern(s) will join a team of researchers that includes two faculty members
Corey Jackson, Ph.D., in the Information School, and Kaiping Chen in Life Sciences Communication. Both have expertise in computational social science/data science and some experience developing applications with RShiny. Their team has worked on developing KM for 1.5 years with a team of graduate students. The UI and backend development has been led by Shelcia David Raj (an experienced developer in the Information School’s Information Science program). However, the project would benefit from additional contributors to lessen the burden on Shelcia. The student could join weekly research group meetings or receive mentorship as needed through in-person meetings or Zoom calls.
Desired Skills:
RShiny/R, GitHub, Open Source
Application Requirements:
Review the available projects by visiting the various project pages.
Interns should apply through the UW Student Jobs portal. Applicants who are not currently admitted or enrolled as a UW-Madison Student or without a UW NetID can login as well as create an account. Please note, you must apply to each project individually that you want to be considered for.
Application materials submitted through the UW Student Jobs portal should include:
A one-page cover letter that highlights your qualifications based on skills identified in the project listing and your interest in open source broadly.
A resume that includes your name, school email address, phone number, field(s) of study (major, minor, degree, certificate), relevant coursework, extracurricular activities, expected graduation date, relevant sample work (ex: GitHub link, personal website, etc.) and any relevant work or research experience.
The names and contact information of three references.
Submit a resume, cover letter, and three references as part of your application.
Interviews will be arranged for selected candidates on a rolling basis after applications close.
