Data Days CLE 2026 (Conference Day)

Property Insights is a dataset and application that contains over 200 attributes ranging across a variety of topics such as building & housing, city planning, public health, fiscal status, and much more! In this presentation I will overview how the Property Insights sausage is made. I'll talk about the data engineering and analytical techniques we used to produce the dataset, alongside a demonstration of how we built the application using ArcGIS Experience Builder.

Your organization runs on data — it's how you know what's working and how you make the case to sustain and grow your programs. But most teams hit a wall after a spreadsheet or two: sources multiply, staying current becomes a chore, and trusting the numbers gets harder just as complexity increases.
The good news: you don't need a big budget or an IT department to fix this. Using free and open-source tools — DuckDB, dbt, and the spreadsheets you already trust — you can build a real data warehouse that runs on your own laptop.
Drawing on 13 years experience in the non profit sector, and first hand experience building data warehouses, this talk walks through what's achievable in 2026: centralizing scattered data, encoding your team's institutional knowledge into clean and tested data, and freeing up time to actually ask the questions you've never had time for. Live demos will show these ideas in action — mapping civic open data with DuckDB, Leaflet, and H3, running real SQL queries, and automating it all with GitHub Actions.
If you've ever wished your data could do more than sit in a spreadsheet, this session will show you exactly how to get started — no procurement request required.

Historical cemetery records are rich in local history but difficult to use as data. This session follows the development of a local, open-source pipeline that combines image enhancement, OCR, vision-language models, and human review to turn scanned forms with handwriting, changing layouts, and imperfect text into structured, searchable records while preserving uncertainty and provenance.

What good is data if it doesn't tell you anything? Cleveland Hearing & Speech Center had plenty of data and no insight, until we built Sound Insights: near real-time visibility for faster decisions and earlier insight than financial statements allow. One example: data showed our clients weren't coming from underserved neighborhoods, so we moved our office there instead. Come hear how we built a data science function, and walk away with steps to turn your own data into strategy.

We are building a new ETL system at The Cleveland Museum of Natural History in order to answer questions that can't be approached with the built-in analytics of our business systems. This talk walks through the system build-out: from the initial motivating questions, to ETL concepts, to the pipelines in production today. The stack is self-hosted on local VMs, but the same principles apply to cloud-native environments. Tools covered: Airflow, Python/pandas, Postgres, SQL, and dbt.