List of Leaders of League of Legends
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
cloud
public
.gitignore
LICENSE
README.md
champ-score-comments.py
riot.txt

README.md

P.E.T.S.

People for the Ethical Treatment of Summoners

We are not your "Minions", we are not your "Monsters". We are living breathing sentients that to live with you in harmony. Rito promised that summoners would FINALLY be NURFed, but it was all a cruel joke!

What the site basically does is expose the champions, items, and regions/tiers, that are the most CRUEL to sentients and the ones that are the NICEST to sentients. Behind the scenes, there's a lot more stats you can sort on though thanks to the API's wealth of data. We processed every bit of non-timeline information that made sense, like which champion dealt the most crowd control or who counter-jungled the most.

In general we chose free-tier eligible platforms for hosting, so we were somewhat limited in computational power and capabilities, so pardon things like non-cached data or slower response times.

Features

Background Jobs

Background jobs were used to process the data as it came in, and update the database with the moving averages.

Responsive Web Design

Our splash page is extremely responsive and is browsable on smartphones.

Merging Data

We merge our content metadata from Parse, statistical data from Parse (and potentially CouchDB), and static data from the Riot API on the fly to build our views.

Advanced Tables

By integrating our data with datatables on the frontend, we did a lot of setup but only barely scratched the surface of what's possible with their API.

Incomplete Features

Unfortunately URF ended too soon, and the timing made it tough for us to spend the time this project deserved.

Map Reduce

As we started exploring the database more we wanted to perform analysis that wasn't capable with a simple relational database let alone moving averages. We uploaded all 19GB of our persisted matches data to IBM's Cloudant to start using a MapReduce view, and were even taking advantage of the Chained MapReduce features of Cloudant!

Leaderboards

Our original plan was to do a leaderboard based on match data beyond what other services do with just your ladder ranking. Leaderboards really lend themselves to using Redis as a backend (because of ZSCORE), but the free tier options on most Redis hosts didn't cut it and there wasn't enough and migrating to a free tier PaaS with Redis support like Heroku or Beanstalk.