To echo and expand on the above grumbles...
Please, go and find a tame maths/data analysis/stats student and go and ask them about data gathering. Morningsider was kind with 'lacklustre' - it's bad enough to utterly invalidate any conclusions you try and draw from the data. Admittedly given the general dodgy standard of data analysis that supports most government decisions the politicians probably won't care, but it would be nice if the effort you put into a masters results in useful information*
As a suggested alternative, why don't you meet some of the folk on this forum (search for PY threads) and talk through what you mean by 'Considerate Cycling'. You'll then get a feel for what a bunch of cyclists understand by the term, and you'll hopefully be able to find a better way to gather data.
You'll also get less bad karma from using a friendly community as a free/easy data source without contributing anything back.
Sorry; it's been a bad day at work, and bad statistics really hacks me off!
*For example, I can confidently say that based on a two dimensional SIR model, the Black Death does not appear to have been a bubonic plague pandemic. Admittedly I have yet to find an actual use for this nugget...