Data scraping to create UK dataset on "use of social media, by education and district"

The data I am in is: use of social media, by education, in the UK districts, from 2008 to 2015. More specifically: - by "use", we mean any proxy either of the extensive margin of use (e.g. number of users) or of the intensive margin of use (e.g. hour spent using the social media / number of log in / posts / actions), or any other proxy you might think at, describing the use of social networks; - "social media", also here we can be flexible, therefore any proxy will be good, e.g. only Facebook would be ideal, but "social media" in general is fine as well, or Facebook and Twitter, etc; - "by education" means any kind of clustering of education level (e.g. college degree / not college degree; or different scales of education), of course the more detailed the better, but also a dichotomous variable could be good; - "UK district" would be the optimal geographical level, but again also something more detailed would be even better; if there is nothing at the level of districts, what would be the closes level of geographical disaggregation you could get? - "from 2008 to 2015", the time period is flexible, therefore any periods or single years you have information on would be good; yearly information would be enough. As it should be clear from the details above, our request is very specific (use, Facebook, by education, UK districts), but can be easily changed in something less detailed (users, social media, UK districts) or even less.