Using my system of * (poorest) and xx (most well off):
Gipsy Hill 88.36%
* Tulse Hill 82.69%
xx Clapham Town 86.94%
* Larkhall 82.28%
Knights Hill 81.88%
xx St Leonards 78.66%
* Brixton Hill 77.90%
Herne Hill 77.79%
xx Streatham Wells 76.66%
xx Clapham Common 73.00%
xx Streatham Hill 70.74%
* Stockwell 68.33%
* Coldharbour 63.72%
xx Thurlow Park 63.65%
xx Thornton 63.49%
Bishops 61.87%
Princes 61.83%
Oval 61.42%
xx Streatham South 60.25%
* Vassall 53.64%
* Ferndale 52.65%
I don't think these figures show any clear connection between whatever it is that they are actually measuring and how deprived a ward is.
You have to bear in mind that working out what percentage of people are registered is difficult and all figures should be used with caution - this is because people who don't register to vote may also not return information to the census either, so comparing the two may not give any decent figures.
It is also not that helpful to simply count the number of properties that have noone regustered to vote, because there is not real way of knowing how many people actually live at each address.
From my own experience of door-to-door canvassing on Railton Road, using the electoral roll, I found whole blocks of properties where there was noone listed at all - or rather the houses numbers didn't even appear on the electoral roll! I also found a fair number of other inaccuracies.
Due to utter incompetance by Lambeth I even managed to run as a (Green Party) candidate but wasn't put on the voters register (yeah I can't work this one out either!

I think they simply used old lists and didn't bother adding my name onto the list they sent to my local polling station), so I wasn't allowed to vote for myself!
The whole thing is a total joke in terms of representing "the electorate", although having said that, if people actually want to register to vote it is very easy to fill in the form or contact the local registrar - as long as you do it far enough in advance of an election (as I found out to my cost).