because the most common number of votes was in between 1 and 2 hundred, the ward sizes are all somewhat consistent so this is a pretty bad application of benfords and a waste of energy. I ran the Clinton numbers from 16 and they showed the same thing, but when you look at the vote counts it makes sense, the democrats tend to get 4-600 votes per ward while trump got about 1-200 in each. Benfords should only be used when you have a mix of values with different orders of magnitude: ie some in the 10's, some in the 1000's and some in the 100,000,000's.