This project analyzed thousands of U.S. public sector job postings scraped from the USAJOBS public API to identify roles that are likely engaged in purchasing or using third-party data. Many such roles are not explicitly labeled as “data buyer” positions, but contain textual and structural signals indicating involvement with vendor platforms, data subscriptions, analytics integration, or fraud detection tools. To uncover these latent buyer roles, the group trained a logistic regression classification model using TF-IDF features derived from job titles, descriptions, and key duties, alongside structured metadata like agency size, industry classification, and role seniority.
The Blue Street Data Capstone Team included:
Kelsey Pohlig, Rory Quinlan, Nachu Lakshmanan, and Zach Vanek