MtGDS, or Magic Data Science, started out of a desire to learn the history and structure of Magic: The Gathering using a quantitative approach. Basically, I kept having ideas, and there is so much Magic data out there, I felt I had to start writing about it.
If you’re new here, I think the most interesting posts I’ve published thus far are this one, about +1/+1 and other creature-enhancing counters, and the one in which I try to understand card design by comparing real cards to random cards.
I have plans to do some more sophisticated modeling of card design, EDH/Commander, Limited, and more. I also have ideas for visualizations and statistics that I haven’t really seen elsewhere. If you want to keep up with everything I write, follow me on Twitter @MtG_DS or subscribe via email below. If you have ideas or feedback, @MtG_DS is your best bet.
Data sources
The foundation of all of my analysis is the MTGJSON AllPrintings file, which I wrangle pretty substantially and use in almost everything you see on this site.
I also make heavy use of Scryfall (obviously?), including their incredible, collaboratively-produced tags.
Most of my work on EDH/Commander uses the fantastic resource that is EDHREC, which actually does a lot of work to aggregate and harmonize decks from across the internet.
17Lands collects data on Limited play on Arena, and shares both summary statistics and large public datasets.
Finally, CubeCobra is my source for anything cube-related. Especially their very functional search tool.
Glossary
I’ve invented or borrowed a fair number of terms, so I’ve written up a Glossary to clarify things a bit!
mtgds 
