What is all this data?
The Idea
So this week I saw a cool video on youtube about how to get data with removed bias from the Riot Games API. The video in question Later in the video it event showed that this is now used on coachless.
Somewhat unrelated I also then saw caedrel on stream “release” his website dpm.lol. And I noticed that these websites are getting very good, and since I’ve always wanted to make something similar I thought I might was well start now.
The Exploration
I instantly started to look at the Riot Games API and found that there is a lot of data available.
I looked through some request and saw that the matches seem to be autoincremeting per region e.g. EUW1_7099405891. Which would mean that there are over 7 billion matches just on the EUW servers.
Combining this finding with the available data that there are about 30 million league players playing on EUW we can assume that there are about 233 matches per player. This then would mean that on average every player plays about 100 hours of league of legends(I know, I know this number is very skewed, but it’s a fun thought).
Then I looked at the data that is available for each match and found that there is a lot of data available. And by a lot I mean a lot.
Each match has the match itself and then a timeline. The match request returns about 100kb of data, but the timeline request returns about 1.5mb of data. And this is just for one match.
Extrapulating this data to the 7 billion matches we get about 10649TB of data(this can change a lot since this depends on game length and the amount of events for each game).
Which got me thinking, what could I do with this data?
The Plan
The plan is to make a website that collects all the information about you, and then builds you an individual profile about where your strong and weak points are.
This could be done by useing the new challenges data, and combining it with the timeline.
Then later I would like to make an item builder that finds the best items in each situation for you. Something similar to coachless.
Another reason I wanted to make this is that each website has a “match score” like “op.gg-score” and “mvp” with ranking, and I’ve noticed that this is very biased. Along this xkcd standards meme line of thinking
Other
There was no other progress this week, as I was occupied with other stuff. So here are some cool videos I found.