This is a story of how I used Instructor and OpenAI APIs to enrich public data and build rocklandcounty.info. This is a small example of how LLMs unlo

Enriching Public Data with LLMs

submited by

Style Pass

2024-04-02 23:30:02

This is a story of how I used Instructor and OpenAI APIs to enrich public data and build rocklandcounty.info. This is a small example of how LLMs unlock new ways to explore and enrich public data.

My wife and I recently bought a home in Rockland County, NY – a county of 350,000 people located twenty miles north of New York City on the west side of the Hudson. The house was built in the 1960s and is need of some updating. Early in our renovation process we learned that any home improvement contractor doing work in Rockland must be licensed by the county – it's Rockland's way of keeping labor local while sitting next to NYC.

The county publishes a list of all licensed home contractors, but it's not particularly good. Search and filters are limited. Data is minimal and in ALL CAPS. Contact information is strictly analog.

As a software developer in the market to hire numerous home contractors over the next few months, this was too tempting of a dataset to pass up. It's small enough to wrap my head around– about 2,700 contractors – but too big to finesse manually.