ACS Data Users Group

How would i go about finding neighborhood demographics data for the neighborhoods I lived in during 1990 and 2000? I would prefer something more detailed than city data. I want to split the data down to the level of the few surrounding neighborhoods.

The closest to a "neighborhood" might be an area known as a census tract. You can look up the census tract IDs for the neighborhood you're interested in using the Reference maps found here: https://www.census.gov/geographies/reference-maps/2000/geo/2000-census-tract-maps.html for 2000 and here https://www.census.gov/geographies/reference-maps/1990/geo/1990-census-tract-maps.html for 1990.

Then look for historical data for those tracts from the 1990 Census and 2000 Census. https://www.census.gov/main/www/cen2000.html

Another option might be to contact your local State Census Data Center for help finding the correct data: https://www.census.gov/about/partners/sdc/member-network.html

In my experience, tracts are a very poor substitute for neighborhoods. I was recently tasked with comparing both tracts and ZCTAs to designated neighborhoods in St. Louis. The takeaway: Tracts are better than ZCTAs, but neither one is a good or even "close-enough" fit. Your mileage may vary of course.

(tracts green, neighborhoods red)

Also, I believe that in 1990, tracts did not have full coverage. AMC8's city might not even have tracts for 1990.

The U.S. was fully covered by tracts for the first time in 1990. (Technically, many of these were "block numbering areas" (BNAs), nominally not census tracts, but the BNA data are delivered exactly as tract data, so it's effectively full coverage.) There's also nationwide coverage of block groups and blocks in 1990, so identifying sets that match up with neighborhoods should be do-able. All of these data are freely available via IPUMS NHGIS.

Yes. 1980 was the last time it was incomplete

You could use the Find Address tool at the US Census Bureau

https://geocoding.geo.census.gov/geocoder/geographies/address?form

It's possible tracts will be good enough if they are not greatly changed in that time. If you want maximum precision, you can use mapping to overlay blocks from that tract onto a street map and pick an exact boundary you think of as a neighborhood, and then you can use that custom shape to "clip" the component blocks from earlier decades, Race/ethnicity and some age (over and under 18) will be there at the block level for each PL94 release after census (it will be very late this year of course),

Given that I suppose you're not very familiar with Census data, I would use something like Census Reporter. They provide a lot of information on the latest data and make it easy to find which Census geography you're in (tract, block group, etc.). censusreporter.org/.../

Thanks for recommending Census Reporter, but it's important to note that we don't serve any historic Census data -- only the current American Community Survey. NHGIS is almost definitely the best source for historic data and maps for this.

I was under the impression they were interested in current information for past neighborhoods they lived in, but right historic data is tricky.

I guess another option is the Longitudinal Tract Database. Look up the tract using Census Reporter (or whatever tool) and then pull the tract out of the LTDB. https://s4.ad.brown.edu/WebGISnew/webgisLTDB/

Caution: LTDB uses a simplistic area weighting method to allocate historical tract data to 2010 tracts. For a more accurate alternative, in addition to providing the original 1990 and 2000 census data, NHGIS also provides 1990 and 2000 data for 2010 tracts (and for 2010 block groups, ZCTAs, and other areas) in standardized time series tables.

(LTDB interpolates historical data from whole tracts, requiring lots of error-prone disaggregation. NHGIS interpolates historical data from blocks, which gives exact counts or near-exact in many cases where LTDB's numbers are tract-based estimates.)