Problem
A local community had a growing list of member businesses in [PLACEHOLDER: the messy source, e.g. a spreadsheet] that was hard to search and easy to let go stale. They needed something usable in the room during a monthly meeting.
Approach
I put the effort where it mattered, in ingestion and cleaning, then wrapped a light Streamlit interface around the clean data. The app is deliberately simple. The reliability comes from the pipeline behind it.
What I built
- A repeatable ingestion and cleaning step that turns inconsistent records into a consistent dataset.
- A Streamlit directory people can search and filter live.
Outcome
Used at monthly meetings by [PLACEHOLDER: who and how many]. [PLACEHOLDER: any other concrete result.]
Why it is here
It is a small, honest example of the unglamorous half of data work: the cleanup that makes everything downstream trustworthy.