Data exploring & Manipulation Project
Simple Data cleaning & Manipulation project
Using a Dataset from Kaggle:
League of Legends Ranked Matches
Data about 184070 League of Legends ranked solo games, spanning across several years and after cleaning 148638 games.
Content:
- Matches.
- Player and team stats.
Okay! let’s go on a quick tour on what’s the Notebook about:
Introduction
- this project is about Data Cleaning & Manipulation using Pandas from python library and Data visualizations using seaborn and matplotlib libraries.
- a dataset from Kaggle was used, {League of Legends Ranked} Matches access it from HERE.
- Exploratory Data Analysis was implemented to find some correlations to gain more insight about some pick strategies.
Doing some Exploratory Data Analysis(EDA):
previous heatmap included only first 25mins of the game and so correlations here are:
- kills & deaths affect strongly the winning process.
- also assits & turretkills affect the winning process.
- kills has strong relation with goldspent.
- more goldspent means more totdamagetochamp means more likely to earn kills.
__
So in the late game as gamers call it OR after 40 mins of game time we find that:
- deaths & kills doesnt even matter alot and have very poor correlation with game winning.
- inhibkills & turretkills have about 25% chance of winning the game(still not big correlation).
- kills have high correlation with goldspent & totdmgtochamp.
- assists have 40% corr with wardsplaced ( as this is the support’s job) also -43% with totminionkilled( supports don’t farm alot) and -32% with kills.
__
Playing with pandas to get top & least win rate champions in each role and printing the results looks like this:
__
Time for some Counter picks advices!
After creating a Dataframe for each role(lane) picks, doing it in seaborn’s barplot and it was huge as there was alot of binary picks. So showing here some of the big differences in each role(lane) as following:
for TOP lane:
__
for JUNGLE lane:
__
for MID lane:
__
for AD CARRY lane:
__
for SUPPORT lane:
__