Modernizing the Data Science Toolkit of a 40-year-old Market Research Company
summary
Over the last 18 months the Decision Sciences team migrates core work to GitHub, producing 200+ repositories, adding code review and project-tracking alongside code, and building internal R packages that speed development; deployment shifts to Posit Connect and Workbench reduce IT burden and stabilize environments. Quarto-based documentation becomes living docs with version control. A planned rewrite of a mature survey-to-report pipeline is deferred after risk assessment. Ongoing training and intermediate GitHub adoption are called out as the main remaining work.
AI-written summary · about
chapters (8)
- 0:00Modernizing matters
- 2:14Modernization at KS&R
- 7:40Migrating to GitHub
- 9:25Organizational change
- 12:01Building the crew
- 13:48IT collaboration
- 15:41Impact and next steps
- 16:06Q&A
related talks
38mA How-to guide for migrating legacy data applications
59mModernize Software and Tackle Tech Debt at Big Bank Scale
20mData-as-a-product: A framework for collaborative data wrangling
24mMake data central feature
15mPackages and Process
25mRepo Resuscitation: Reviving Neglected Projects
36mHow Research Teams Can Deliver Higher-Quality Insights Faster
26mWorking with Data Scientists and why I left my job (not)