1,000 matches were processed to generate this data. In each game, a hero was considered laning against an opponent if that opponent was in either a solo or duo lane against that hero. This gave about 10,000 matchups for analysis.
It's important to note that none of this data proves one hero is "better" than another at laning. There are a number of confounding variables. For new heroes, players will get better over time as they learn their ins and outs. For all heroes, individual builds can have a drastic impact on the outcome of the game (e.g. spirit vs gun build). Some heroes simply have a much higher skill ceiling than others. We aren't splitting out by rank. Additionally, we are bound by the limitations Deadlock places on us; there just aren't that many pocket / mirage matchups, for example. There are obviously many other factors, and it's simply not feasible to control for all of them.
And because no matter how many times I leave a caveat on a project like this, if someone decides they still want to be annoying about this in the comments, feel free to paste the following:
That's such a good point, and you are the first to have thought of it! Someone should tell the author. Maybe he can put a notice at the bottom of the page. I wonder if anyone will read it...
Special thanks to deadlock-api.com for providing some of the necessary data to do this project. If you run one of those deadlock stats sites and are interested in getting this built out and updated automatically on your site, let me know (ellisd23 at the mail site for google).