Use this REAMDE.md file to describe your final project (as detailed on Canvas).
Names: Wenyi Sun, Kai Daniels, Rajeev Vijayaraghavan, Michael Gov
Why are you interested in this field/domain?
Our group is primarily interested in video games and online culture, and League of Legends is one of the largest games in the world right now and has a significant amount of data available about it. As a result, we thought it would be really cool to do our project on such an iconic game. Also, all of us have played video games and felt that League has been culturally relevant for over a decade now, so that also contributed to our interest.
What other examples of data driven project have you found related to this domain (share at least 3)?
- User specific information regarding a players stats, winrates, champion play rates: Mobalytics
- Champion (in game character) statistics regarding itemization paths, win rates, matchups, etc: op.gg
- League of Legends statistical report on demographics, revenues, and e-sports: Project
What data-driven questions do you hope to answer about this domain (share at least 3)?
- Which champions (characters) have the most popularity and what are their corresponding win-rates?
- How winning certain objectives in game affect your probability to win? (For example, getting the first kill or first tower)
- Does champion selection affect game duration and win rate? Do some champions tend to have longer game durations?
Where did you download the data (e.g., a web URL)? How was the data collected or generated? Make sure to explain who collected the data (not necessarily the same people that host the data), and who or what the data is about? How many observations (rows) are in your data? How many features (columns) are in the data? What questions (from above) can be answered using the data in this dataset?
First Dataset:
Description: Professional players for League of Legends Championship statistics
Download link: Champ stats
How was the data collected: The data was collected by Oracle's Elixir, a blog dedicated to recording stats of League of Legend's esports scene.
How many observations (rows): There are 1428 observations
How many features (columns): There are 98 total features
Which questions from above can be answered using the data in the dataset: Which champions are the most popularity and their win rates, as the esports data would contribute heavily in understanding win rates of champions and their popularity.
Second Dataset:
Description: Massive EUW Dataset of League information
Download link: League Data
How was the data collected: This is a collection of over 50,000 ranked games of League of Legends from their West European servers and takes into affect all parts of a single League of Legends game.
How many observations (rows): There are 51,490 observations (Games)
How many features (columns): There are 61 features
Which questions from above can be answered using the data in the dataset: This can answer how winning certain objectives in game affect your probability to win.
Third Dataset:
Description: League of Legends champion data
Download link:Champ Data
How was the data collected: The data was collected from the public League of Legends API by Riot Games.
How many observations (rows): There are 145 observations (champions)
How many features (columns): There are 11 features.
Which questions from above can be answered using the data in the dataset: It answers the question on how champion selection affect game duration and win rate, as it considers every champion's invididual ability.