Reflections on a year of using generative AI tools in community research — what works, what keeps me up at night about data governance, and the questions we need to answer as an institute before we go further.
Why bilingual fluency in R and Python is becoming essential for community research teams
After years of defaulting to ArcGIS, a real project finally pushed me to give QGIS a serious try and I never went back. Here is what changed, what QGIS does better, and how R has quietly taken on a meaningful share of that workload too.
Bayesian Improved Surname Geocoding (BISG) is a probabilistic method that estimates racial and ethnic group membership using last names and residential geography — a practical tool for working with administrative data that lacks race and ethnicity fields.
The American Community Survey Public Use Microdata Sample (PUMS) gives you individual-level Census records for custom cross-tabulations and analysis. This tutorial walks through using the tidycensus package to pull and analyze PUMS data for Mecklenburg County, NC.
Introducing elements of data pipelines used in community research at the Urban Institute
reconciling our work with the strategic goals and movements at the institute
Dealing with the legacy mess and finding new footings
On developing a reproducible data science framework and practice at the Charlotte Urban Insitute
Improving Reproducibility in Our Research Workflow
Building standard documentation for analytical research Practice at the Urban Institute
On how we developed the technical operations manual at the Charlotte Regional Data Trust
Developing and automating metadata management at CRDT