How do you make this? It doesn't seem to be like Wikipedia has coordinates or map boundaries for ancient empires, so there's no simple way to mine the data.
And if you don't mine it from somewhere, how do you know what to include? How many people will have heard enough about every part of the world to even be able to research ancient borders?
This is the same question as "How would the information even get put into something like Wikipedia in the first place" - knowledgeable people in the field do the work of aggregating academic information and turn it into a database of sorts. This seems to be accomplished in this case by running it as a business with a founder who has a degree in geography and history + a team of variously skilled people (history + tech + business).
It's a very fine-grained data, isn't it... Something like on August 24, 1652, 17:34 local, soldier X killed the last tribesman Y they were fighting and claimed the area for his king Z. And a lot of regions were/are disputed with no clear "ruler", e.g. the no man's land between North and South Korea.
I guess something like this - add timelines for every known point (city, landmark, ritual site etc.) connect the dots in the same year with one another, then apply some reasonable territory estimate to round out the resulting blob, correct for visible mistakes manually.
Just for fun, how would you envision this process working for empires that weren't based on territorial control of cities and landmarks like the Mongols, or based on territorial control of land at all like the Hansa?
I guess the same way, based on the known extent of the empire, put some points which they held the same way as cities, and then go from there. More manual work, but the same idea. This is just guessing btw.
And if you don't mine it from somewhere, how do you know what to include? How many people will have heard enough about every part of the world to even be able to research ancient borders?