Lets look at some numbers. The Atlanta Journal Constitution reports the results by city council district (and also precinct) here and also provides an excellent interactive map of the city. Their data also includes each district's white and black population %. From this I created the following chart:
What we see is that the voting does not show a consistent correlation between the racial makeup of the electorate and its vote. If we were to hypothesize an electorate perfectly polarized by race, we'd expect each candidate's performance to essentially match the racial composition of each district. That is far from what happened. In the three whitest districts (6, 7, and 8) Reed did better than we might expect. Likewise, in the four blackest districts (4, 10, 11, and 12) Norwood also outperformed expectations based solely on race. These numbers are even more interesting when we throw the variable of party into the mix. We would assume that those whitest parts of Atlanta, located in the northern part of the city, would be the most Republican. That he still reasonably well here is quite fascinating. Likewise, although Norwood campaigned as an Independent, she has more often than not voted as a Republican in past elections. That she could perform so well in heavily black areas further suggests that there were some interesting dynamics at play.
When the Civil Rights Movement threw much of the south into turmoil, there was a saying that Atlanta--which didn't produce the violence seen in places like Birmingham, Selma, and Oxford--was a "City Too Busy To Hate." Maybe yesterday's vote was an indication that many people in Dixie have moved on.