The Redistricting files are being released on a state by state basis by the Census Bureau.  These files contain total population by race and latino origin, and population over the age of 18 by race and latino origin.  The website is: http://www.census.gov/clo/www/redistricting.html. The data are released at block level.  However, the geographic header includes a TAZ code, so that we can aggregate the data to a TAZ level.  I tried to download the data, transfer them to a GIS, and aggregate the data to a TAZ level for a county, and state.   I worked on Windows NT, however the process should work on Windows 95, 98, or 2000.  For unix users, I have a small note at the bottom. Please note that this is *a* method.  There may be far better methods than this.  If you know of any other methods, or have written code to automate the transfer, please let us know.
 
Here are some general steps in a PC environment:
 
There are three files that need to be FTPed from the site for each state.  I attempted to convert their files to a GIS Platform, specifically Arcview. 
1. For any county or state download the shape file for block data from ESRI's website (http://www.geographynetwork.com/data/tiger2000/).  Unzip these files, and keep them in a specially created folder.  I created a folder called d:\VA_PL_Data, and saved the extracted shape files as 2000blocks.shp.
 
2. FTP the three redistricting files from the CB ftp site at http://www2.census.gov/census_2000/datasets/redistricting_file--pl_94-171/
 
3. For example, for virginia, the geographic header file is  vageo.upl.zip, and the data files are va00001.upl.zip  and va00002.upl.zip 
 
The files need a bit of massaging before import to Arcview.
Step a: The geographic header file is a flat ASCII file.  The fields most of us are interested in are:
1. The State and County FIPS.
2. The Tract
3. The Block
4. The TAZ
To convert the flat file into something that can be used for linking fields, you will need to know where (or at which character) a record/field begins and ends.  This can be obtained in chapter 2 (pages 16,17,18,19,20,21,22) of the file layout document (http://www.census.gov/prod/www/abs/pl94-171.pdf). 
For example:
You can open the file in MS Access and specify exactly the 18th marker to the 25 marker for the "logical record" which is the key between the geographic header file and the data files.
The State and county FIPS are between the 29 and 34 markers.
The Tract is between the 55 and 61 marker. (Meaning that the record begins at 56 and is of 6 characters, so it ends at 61).
The block is between the 63 marker and the 67 marker.
The TAZ is between the 347th marker and the 353 marker.
 
Once the table is imported into Access, save it as a dbf file for import into Arcview.  I called the geography file as va_geography.dbf and saved it in the folder d:\VA_PL_Data
 
Step b.
The data files are fairly straightforward to convert as they are comma delimited.
va00001.upl.zip:  The file contains total population by race, and hispanic origin, and is in a convenient .csv format, and can be easily imported to access and converted to a dbf.  The first field in this file is the "logical record key" - a field you can use for connecting the data with the geographic dbase file developed in step a. I named the data file as va_totalcounts.dbf and saved it in and saved it in the folder d:\VA_PL_Data.
 
va00002.upl.zip contains voting population (Age > 18) and is formatted in the same way as va00001.upl.zip.
 
You can import these into access and give them some headers that are more recognizable for you.  Export these as DBF too.
 
Step 4: 
a. Open a new Arcview session, and add the block theme you downloaded in step 1 (in my case 2000blocks.shp) as a theme.  Also, add the three tables developed in Step 3.  You will notice you can "join" the geographic header file and the data files easily in Arcview.  I needed only total population counts by race, so I had only two tables to join (va_geography.dbf, and va_totalcounts.dbf).  I used "logical record field" as the key field to join the two tables.
 
 
b. However, the block theme obtained in step 1 (2000blocks.shp) has an attribute table where the state fips, county fips, tract, and block number are concatenated into a single field.   So to join the shape file to the data file, we need to concatenate the fields containing the statefips, countyfips, tract, and block in the new va_totalcounts.pdf obtained in step 4a.  This can be then connected to the corresponding field in "Attributes of 2000blocks.shp"
 
Once the data are in Arcview, you can turn on the geo-processing wizard and summarzie the data by TAZ, or by Tract easily.
File- extensions- click on geoprocessing wizard.
Then go to view and dissolve the block shape file on TAZ, Tract or anything else.  This is the reason I added TAZ as an import item in step 3.  You can add other fields you are interested in summation, or geography. 
 
UNIX Users:
You can follow most of the steps outlined for PC users.  You may have to save your files as comma delimited and import them as .txt files into Arcview.  The steps that used Microsoft Access are basically far easier in a UNIX environment.  For example, for the geographic header file, you can use shell scripts or simple "cut" and "paste" commands specifying the exact start and end points for fields (obtained from the file layout manual).  The rest of the process must work.
 
Thank you!
Nanda Srinivasan
202-366-5021