A ‘cluster’ of cases? One of the reasons the ZOE study was created was to help identify local clusters before the official data (or because the official testing didn’t work when ZOE was set up). Was there an outbreak at a factory or schools maybe? At a local level it doesn’t take ‘that’ many cases to cause a spike. In all the time I’ve used that app though the Birmingham data has never had such a sudden spike, even when the Unis returned and there were loads of cases , but maybe it’s larger population ‘hides’ a cluster to some degree.
FWIW I don’t think they looked at that level of detail, they took the rate across much bigger areas, and across the Midlands as a whole the rate was pretty bad. We’re all in it together remember