Big commit here.
This moves the "data" directory inside the website, and then uses fetch to bring the csv and geoJSON files into the map instead of copy-pasting into JS files. Much cleaner and reduces steps from updating a CSV to updating the actual website. Now we simply need to git pull on the webserver (unless the updated data is for library services rather than library locations).
Also cleans up some styling.
Also removes my MapBox secrets from the Git repo whoops. These are now brought in via a secrets.js file that simply declares a constant. This file will not be visible in the repo.
Fixes#14
Hugh Rundle
