Project that uses PRAW, the Reddit API wrapper, to retrieve and analyze data from r/leagueoflegends. In the code, I retrieve the top 150 posts from the subreddit, then crawl/clean through all post details, comments, and replies. Then, I then fuzzy match on champion names to obtain data on how often a champion is mentioned.
Raw results:
[('yasuo', [235]), ('fiddlesticks', [227]), ('brand', [162]), ('teemo', [159]), ('thresh', [146]), ('vayne', [128]), ('graves', [116]), ('singed', [110]), ('lux', [109]), ('ryze', [108]), ('skarner', [105]), ('garen', [99]), ('akali', [97]), ("kai'sa", [95]), ('sion', [92]), ('dr. mundo', [86]), ('galio', [86]), ('xayah', [82]), ('mordekaiser', [80]), ('nasus', [80]), ('pyke', [78]), ('aatrox', [74]), ('urgot', [74]), ('nunu and willump', [71]), ('aphelios', [70]), ('heimerdinger', [70]), ('lucian', [70]), ('jhin', [69]), ('amumu', [66]), ('gragas', [66]), ('ezreal', [62]), ('riven', [62]), ('zilean', [62]), ('senna', [61]), ('irelia', [58]), ('zoe', [58]), ('sejuani', [57]), ('ahri', [56]), ('ivern', [56]), ('jinx', [56]), ('shaco', [56]), ('sylas', [54]), ('swain', [53]), ('malphite', [52]), ('veigar', [52]), ('draven', [51]), ('neeko', [50]), ('yuumi', [50]), ('leblanc', [49]), ('rakan', [49]), ('varus', [49]), ('zed', [48]), ('bard', [47]), ('tristana', [47]), ('syndra', [46]), ('leona', [45]), ('sett', [45]), ('braum', [43]), ('alistar', [42]), ('kassadin', [42]), ('warwick', [42]), ('ashe', [41]), ('darius', [41]), ('kalista', [41]), ('lissandra', [41]), ('lulu', [41]), ('caitlyn', [39]), ('pantheon', [39]), ('taric', [39]), ('twisted fate', [39]), ('camille', [38]), ('janna', [38]), ('nocturne', [38]), ('qiyana', [38]), ('taliyah', [38]), ('olaf', [37]), ('rengar', [37]), ('shen', [37]), ('jax', [36]), ('orianna', [36]), ('renekton', [36]), ('ekko', [35]), ('ornn', [35]), ('fizz', [34]), ('gangplank', [34]), ('jayce', [34]), ('soraka', [34]), ('nautilus', [33]), ('poppy', [32]), ('gnar', [31]), ('annie', [30]), ('blitzcrank', [30]), ('karthus', [30]), ('vi', [30]), ('zac', [30]), ('sivir', [29]), ('anivia', [28]), ('maokai', [28]), ('nidalee', [28]), ('rumble', [28]), ('sona', [28]), ('viktor', [28]), ("rek'sai", [27]), ('morgana', [26]), ('malzahar', [25]), ('tryndamere', [25]), ('jarvan iv', [24]), ('tahm kench', [24]), ('talon', [23]), ('ziggs', [23]), ('azir', [22]), ("kog'maw", [22]), ('wukong', [22]), ('xerath', [22]), ('illaoi', [21]), ('cassiopeia', [20]), ('elise', [20]), ('kayle', [20]), ('kayn', [20]), ('kled', [20]), ('shyvana', [20]), ('udyr', [20]), ('zyra', [20]), ('corki', [19]), ('katarina', [19]), ('kindred', [19]), ('trundle', [19]), ("cho'gath", [18]), ('diana', [17]), ('fiora', [17]), ('kennen', [17]), ('lee sin', [17]), ("vel'koz", [16]), ('quinn', [14]), ('rammus', [13]), ("kha'zix", [12]), ('miss fortune', [12]), ('nami', [12]), ('vladimir', [10]), ('yorick', [9]), ('evelynn', [8]), ('hecarim', [8]), ('karma', [8]), ('volibear', [8]), ('xin zhao', [6]), ('twitch', [3]), ('aurelion sol', [0]), ('master yi', [0])]
Plot:
- Run this script at a cadence on inbound posts. Start goal was to get data on most controversial champions in league history (apparently, Yasuo!), but would love to extend this functionality to create a "trending champions" feed.
- Generalize this script to search any subreddit with an input of a list of names and an output of mentions.
- Champion name matching is hard, even with fuzzy matching.
- Added an adjustment ratio to Karma and Twitch. Karma, since Reddit has a karma system on their posts. I did this by analyzing data on Karma (the champion) vs karma (Reddit system) manually to develop a mention ratio. The data is still really inaccurate, but doesn't matter because nobody cares about the champion Karma. Twitch has a similar problem (twitch.tv – league is regularly streamed there).
- Realizing that I should add more name alternatives in my name dictionary. Example: nobody types "Master Yi" – I should be searching on Yi as well. It'd be even better if I could tighten fuzz.ratio for that name (don't want to match on things like "yee" or "yo").
I got into poetry a while back ago. Here's my attempt at highlighting rhyme schemes, just like Genius does in their infamous videos.
I'm hoping to make this more intelligent so it can start interpreting rhymes within the verse as well.