Analyzing Twitch Gaming Data

by @tylershienlim

Project Overview

About Twitch:

Twitch is the world’s leading video platform and community where millions of people and thousands of interests collide in a beautiful explosion of video games, pop culture, and conversation. Its live and on-demand video platform forms the backbone of a distribution network for video game broadcasters including professional players, tournaments, leagues, developers and gaming media organizations.

This project has two objectives:

1. Analyze Twitch Gaming Data (Using SQL)
2. Visualize Twitch Gaming Data (Using Python)

Technologies & Tools used

• Python
• VSCode
• Jupyter Notebook
• SQL
• Microsoft SQL Server Management Studio
• Other dependencies as required.

Prerequisites / Installation

1. Host local server in SSMS using SQL Server Developer
2. Connect SSMS to local hosted SQL Server
3. Load raw data sets (chat.csv & video_play.csv) into SSMS

Alternative

1. Use convert.py to convert datasets into .db files to use SQLite instead of SSMS

Data Preprocessing

Data was preprocessed using Microsoft's SQL Server Management Studio

Code Structure

Below is the tree for the code structure.

1. Data is differentiated into raw and processed data
   • Raw data is the original data sources obtained from Codecademy
   • Processed data are the outputs as a result of the SQL Queries executed
   • SQL Queries can be found in the twitch-exploration.sql file
2. Using the processed data found in ./data/processed directory:
   • Bar graph is visualized
   • Pie chart is visualized
   • Line plot is visualized

📦cc-twitch-gaming-data
 ┣ 📂data
 ┃ ┣ 📂processed
 ┃ ┃ ┣ 📜distinct_channels.csv
 ┃ ┃ ┣ 📜distinct_games.csv
 ┃ ┃ ┣ 📜genre_classification.csv
 ┃ ┃ ┣ 📜LoL_viewer_countries.csv
 ┃ ┃ ┣ 📜popular_games.csv
 ┃ ┃ ┣ 📜streaming_modes.csv
 ┃ ┃ ┣ 📜view_count_hour_total.csv
 ┃ ┃ ┗ 📜view_count_hour_US.csv
 ┃ ┗ 📂raw
 ┃ ┃ ┣ 📜chat.csv
 ┃ ┃ ┗ 📜video_play.csv
 ┣ 📜convert.py
 ┣ 📜README.md
 ┣ 📜twitch-exploration.sql
 ┣ 📜twitch-plots.ipynb
 ┗ 📜twitchlogo.png

Results

3 different visualizations were created.

Acknowledgement / References

This project including datasets were provided and guided by Codecademy part of the Data Scientist: Analyst Course Dataset was provided by Codecademy