Creating Time Series Data from the American Community Survey

I just wrote a blog post about my experience creating a time series dataset from the ACS.

On the surface this sounds easy: Just loop over the years, and add rows row each year.

But I ran into two big surprises with this approach:
- Variables sometimes change meaning over the years
- Geographies sometimes change over the years

My post goes into details about how I solved these problems. And since my code is all open source, others can use and build on my code as well.

I hope it helps someone!