Do you know of a good data source for a US demographics data science project?

[Mr. Morden]

I'm in the middle of trying to make a career transition into data science, and as part of the class I'm taking now, I have to do a ~2 or 3 week project on the topic of my choice.  Was thinking about doing something US demographics-related.  Probably not something overtly political, so nothing on US election results.  But maybe something along the lines of predicting population growth by county based on demographic information from each county.  Doesn't have to be that.  That's just an example.  But I don't know where to find a table with demographic info for every county in the country over time.  (Preferably in an easy to read format like a CSV file.)

So wondering if anyone here has insight on this.  Or if not that specifically, then any similar US demographics-based dataset would be of interest.

[JA]

The American FactFinder is probably your best bet for anything related to demographics data.

The Center for Economic Studies is part of the Census and has a detailed list of microdata sources for a variety of topics.

The Bureau of Labor Statistics also provides comprehensive data for demographic issues pertaining to the labor market.

Here are some demographic projections in this .pdf and a list of different demographic projections at this link.

[Crumpets]

Statisticalatlas.com

[Torie]

One thing I read in college a zillion years that caught my interest, is that when a metro area gets larger than 400,000, it tends to have the infrastructure and offer services, that insure that it will continue to keep growing, and will no longer face the risk of having a substantial population decline. It would be interesting to see how that theory worked out, 50 years later and if it did not, why.

[cinyc]

One thing I read in college a zillion years that caught my interest, is that when a metro area gets larger than 400,000, it tends to have the infrastructure and offer services, that insure that it will continue to keep growing, and will no longer face the risk of having a substantial population decline. It would be interesting to see how that theory worked out, 50 years later and if it did not, why.

Many 400,000+ rust belt metros like Cleveland and Pittsburgh have lost population in this and other recent decades. I'd say no. I've been mapping historic population changes in CSAs and metros here.

[Mr. Morden]

Thanks guys.  I appreciate all of your suggestions.  I've been busy this past week, so have not had a lot of time to think about this, and have only just had a chance to browse a couple of the suggestions you've made.  I will need to narrow down my options quite a bit in the next week though.

One thing that I’ll note is that I find webscraping tedious and I’m also not very good at it.  So to the extent that it’s possible, I’m hoping to be able to use datasets that are conveniently already packaged in some readable format, like CSV or just plain text.  E.g., the county_facts.csv file found here:

https://github.com/Viralshah009/INFM600

is exactly the kind of thing I’m looking for.  Though it would be great to have some of that same info for decades past, and also in an ideal world, from some source other than “some guy on Github”.

[Mr. Morden]

OK, despite the fact that I said I wanted to something more demographics related and not actually political, I have to come up with three different ideas, and one of them will be picked.  So I was thinking maybe for one of my ideas, I could do something with the DW-NOMINATE data on voting similarity between members of Congress.  But I'm not sure if the raw data are available anywhere.  E.g., when I click on the VoteView database link here:

https://voteview.com/about

I get a "404 Not Found" error.  Anyone know if that info is available anywhere these days?