Handling areas with missing mobility data

[wyuill]

Many areas have missing mobility data for some or all days/mobility types. Imperial handeled this by using the last available value but for some areas in the UK there are a large number of missing values potentially making this invalid or even impossible. For now I have excluded areas with 30+ missing values in any column but this doesn't feel like an ideal solution.

Potential soltuions are:

• remove areas with a whole column blank
• remove areas where the starting values are blank
• use some different impute method