One Engineer, an API, and an MVP: Or how I spent one hour improving hiring data at my company.
summary
Builds an MVP in ~1 hour that calls the greenhouse API, pulls jobs and applications, extracts EEOC fields, anonymizes and tallies counts, and renders per-role anonymized EEO aggregates that flag whether a pool meets a chosen diversity threshold. Key outcomes: hacky Django view works but takes 10 minutes live, syncing to a DB fixes performance, two underrepresented finalists notably shift hiring odds, and major gotchas are permissions, API keys, data sparsity, legal limits, and minimum-pool-size anonymity.
AI-written summary · about
chapters (8)
- 0:00Start
- 0:17Nicole Zuckerman
- 1:16Why Is It Hard for Well-Intentioned Organizations To Shift Their Demographics
- 4:26Api Calls
- 9:14Anonymized Our Data
- 10:33Show Results Only for Groups of Minimum Pool Size
- 11:22Importance of Anonymizing Data
- 15:28What Is Sufficient Diversity
related talks
28mPractical steps to establish and scale your data & ML team
34mFrom zero to hero in 6 months
43mBehind the Scenes at Gapminder: A Story about Data and Visualizations
27mEngineers to Data-Scientists with Python
31m"Notes on Over-Engineering: A Project Post-Mortem"
22mGovernment dashboards and open data.
29mHow to make public data more accessible with "baked" data and DuckDB
42mInside Out - Creating a Python Ecosystem for Analysis and Better Government