Following up on my entry on neighborhood "tipping points":
Commenter Nick pointed out that the maps of Austin I posted there use 20% class breaks, which perhaps causes them to exaggerate the degree of segregation. For example, two census tract will have the same shading even if one is 19% African-American and the other 1% African-American. That's a fair point (although the class breaks do a good job of showing census tracts in which African Americans and Hispanics predominate.)
So I generated a couple of new maps from the Census website. They show, for 1990 and 2000, the percentage of African-Americans residing in each block group in central, south and east Austin. (Click the maps to enlarge.)
I generated the 2000 (second) map first and let the Census mapping tool generate natural class breaks, which explains the weird ranges. I then used the same class breaks for the 1990 map.
The maps speak for themselves, but three fairly obvious observations:
1. Things didn't change much west of I-35 between 1990 and 2000. A handful of block groups saw a rise in percentage of African Americans, more saw a decline, and most saw no change.
2. In 2000, African-Americans made up a low (<7.4%) or very low (<3.4%) percentage of most central and south Austin block groups. Central Austin north of Lady Bird Lake had the largest concentration of block groups with a very low percentage of African-Americans.
3. Things changed more dramatically east of I-35. The heavily African-American block groups tended to shift east, away from I-35.
N.B. These maps do not necessarily imply that the absolute population of African-Americans dropped in a particular block group -- the block group simply may have added a lot of residents of other races.
