Tune population density quintiles for Scotland

[michaelkirk]

FIXES #143

Previously we were using density quintiles from the uk/wide census.

We had previously talked about doing something dynamic to scale quintiles per/study area, but I thought that Scotland-wide quintiles was an easier start, and an uncontroversial improvement.

(after on the left, before on the right)

Glasgow:

Dundee:

Angus:

For smaller communities, it's still a bit washed out. Dynamic quintiles would address that.

One downside of dynamic quintiles is that you'd lose the ability to compare density by color across maps, e.g. if you are looking at two maps. My guess is that isn't a huge problem though. Mainly I just like that this approach is simpler, and maybe "good enough".

[dabreegster]

These look fine to me! I could see an argument both for scaling for each study area, and for that being confusing if you're used to the darkest color meaning something very dense in an urban area. Let's go with this for now.