Sept 9, 2014; revised Nov 13, 2014
This page explains how to get:
1) Tract-level Census data for Counties from American Factfinder, and
2) Shapefiles (geospatial data) for the same counties.
It works for the 2000 and 2010 Decennial Censuses, and the rolling-sample data from the American Community Survey. Getting older Decennial Census data is harder. I have posted instructions and some older Census data on another page.
RESET YOUR WEB BROWSER’S DOWNLOAD PREFERENCES. Whichever browser you use to get this data, change your settings/preferences so that the browser always asks you where to download files.
I have revised this page to focus only on finding Table P5 and the Shapefile of your county. You can obtain other Census data files using this same method. Begin by going to American FactFinder:
When you click the Advanced tab, you will go to the Advanced Search page. Here you can filter through the Census Bureau’s databases by setting up search filters. On this website, search filters appear in the box on the left called Your Selections.
Begin creating search filters by clicking on the Geographies tab. A gray-bordered overlay dialog box will appear on the right.
In here, pick Census Tract – 140 as your “geographic type”–meaning your level of data. Then pick your State, then pick your county;
In the Tracts field, select the first option: “All Census Tracts within {yourcounty}, California.” (Note that I use curly baces {} to mark out items where you need to substitute your own item.) Then click ADD TO YOUR SELECTIONS.
Now that you have finished using this gray-overlay GEOGRAPHIES dialog, close it. Since the dialog-box is wide, you may need to pan right in your browser to find the CLOSE button.
Now you will see your Geographies filter in Your Selections on the left, and you can add one more filter by name: P5.
On the right-hand side of the peach-colored “Refine your search results” field, click the GO button and you will see the possible tables below it reduced to only three options:
Pick P5 by clicking on the hyperlink. Now you can view the table itself. Take a look at the data to make sure you are getting Tract-level data. It is possible to pull up Table P5 (or other Census Tables) with only County-level data, in which case you will only get one line of summary numbers for the whole county. Note: in this web-page preview, the table is rotated so that Tracts are shown as columns, and the variables are shown as rows. When you download the data, the CSV file will have Tracts as rows and variables as columns. If you are convinced that you have the data you want, click Download:
The dialog box that pops up has many options. Choose ONLY ONE: Comma-delimited (.csv) format. Uncheck the “Include descriptive data element names” checkbox. If you leave that checked, a second row of data will be added you your file. It is a useful row–because it actually describes the meaning of each data-column–but I want you to keep that data in a separate file called metadata.csv.
I don’t care what your overall directory structure may be, but for this class please create a directory called soc393, with a subdirectory called census. I will call this the path to your files. I will represent it thus: ~/soc393/census/
A brief note about where to place your data
Where should you store the download on your hard drive? Don’t just leave it in /Downloads with a zillion other random files. Move it to a specific subdirectory called:
~/soc393/census/
I don’t care what the overall organization of your hard drive might be. The tilde (~) represents all that: it is an abbreviation-marker for whatever your directory-structure might be above the folders I have just required. In my case the full pathname to my directory is:
/home/pietro/Documents/chron/2014/soc393/census/
Everything above /soc393 will be different on your computer, and that is why I use the tilde to show a truncated pathname.
Okay: now that you have placed the downloaded ZIP-archive in ~/soc393/census/, extract all the files into a new subfolder. You can leave the default name (DEC_10_SF1_P5) or use a more intuitive name. I named it this:
~/soc393/census/Contra2010_table_P5/
It should have at least three files in it:
DEC_10_SF1_P5.csv
DEC_10_SF1_P5.txt
DEC_10_SF1_P5_metadata.csv
The first .CSV file contains the data that we need. The .TXT file is a backgrund description of the information from the Census. The “_metadata.csv” is that ‘separate annotation file’ that was alluded to in the Download Options dialog box. It is an explanatory key to the totally abstract code-names for the variables in the main file. For example, in Table P5, the column “Doo3″ is the number of non-Hispanic Whites residing in each Census Tract. Column “D007″ is non-Hispanic Native Hawai’ians and other Pacific Islanders residing in each Tract. You may even want to print out this metadata file just to keep track of what your data means.
NOTE: You must unzip the downloaded file in order to use it! In Windows, you can look into the contents of a zipped archive automatically, but other software like Excel and R-Studio will not be able to open the files. So unzip the archive! As far as I have seen the default settings on all unzipping software will create a new folder, with the same name as the .zip file, and put the decompressed contents into that folder.
From here, there are several possible next steps:
1. Get geographic (shapefile) data for the same county and tracts;
2. Get similar data for the 2000 Census;
3. Get older data from 1960-1990;
4. Compile the data into a “master data file” which you can use for many types of analysis.