This page is a not so organised collection of datasets I use or want to in the future!
More to follow soon …
OpenCorporates - note that OpenCorporates can grant API keys with generous limits for academic and third sector research.
OpenCage - A paid for API (with R and Python libraries) to improved OpenStreetMap data, well worth the modest subscription fee if you are trying to geo-code serious amounts of data.
EDINA Digimap - An amazing database for UK mapping data including Ordnance Survey, historical, geological, LiDAR and marine maps and spatial data. Most universities and many collages will have access, but unfortunately is not open to all due to copyright issues.
Open Geo-located Internet Speed - Provided by Ookla (CC BY-NC-SA 4.0)