/video-game-encyclopedia

474417 Games with Metacritic Score, Ratings, Genres, Publishers, Platforms, ...

Primary LanguageJupyter NotebookMIT LicenseMIT

Video Game Encyclopedia

Gaming

Visit Kaggle page for the full Game Dataset

Directory Tree

.
|-- src
|   |-- Combine_game_info.ipynb
|   |-- Get_game_id.ipynb
|   `-- Get_game_info.ipynb
|
`-- data
    |-- game_id.csv
    |-- game_info.csv
    |-- game_id
    |   |-- 1.json
    |   |-- 2.json
    |   `-- *.json
    |
    `-- game_info
        |-- 1.json
        |-- 2.json
        `-- *.json

How to Start

pip3 install -r requirements.txt

Steps to replicate the dataset:

  1. Run Get_game_id.ipynb. This makes request to all pages in https://api.rawg.io/api/games?page=1 and save one JSON file for each page in ./data/game_id/*.json where * is the page number. At the end, ./data/game_id.csv is created which contains the name and id of each game which is needed for step 2.
  2. Run Get_game_info.ipynb. Using the id from Step 1, this script makes request to https://api.rawg.io/api/games/ and save one JSON file for each game in ./data/game_info/*.json where * is the game id.
  3. Run Combine_game_info.ipynb. This combines data in ./data/game_info/ and saves it as ./data/game_info.csv. game_info.csv contains the final dataset

Important Notes:

  • File sizes:
    • ./data/game_id with 25000 files has the size of ~15MBs.
    • ./data/game_info with 470000 files has the size of ~230MBs
  • To increase the speed of obtaining the data from RAWG API, concurrent programming is applied to step 1 and 2. However, execution time depends greatly on internet connection speed.
    • Step 1 takes ~1 hour with 32 threads
    • Step 2 takes ~3-5 hours with 64 threads
    • Step 3 takes ~2 minutes
  • RAWG API has a limit of 500,000 page views a month. As more games come out in the future, it would be hard to get all game information in one run without exceeding this limitation.
  • When 1 thread fails while requesting data, it will skip to next game/page automatically. To make sure you get all games from RAWG, you can run Step 1 and Step 2 multiple times. Downloaded files are skipped automatically.

Limitations:

  • To reduce the file size of downloaded files and the final CSV dataset, not all JSON information is downloaded. If you want more customization, you will need to change how the JSON is handled in Step 2
  • Although Multithreading is applied, the whole process can take up to ~6 hours to finish because of the large amount of data.

Context

This dataset contains 474417 video games on over 50 platforms including mobiles. All game information was obtained using Python with RAWG API. This dataset was last updated on Dec 22nd 2020. If you are interested in obtaining more recent games, visit the GitHub page for more information. I plan to update this dataset annually.

Content

Each row contains information about one game. There are several columns that have multiple values like platforms, genres, ... In those cases, values are separated by double pipes ||.

Column definitions:

  • id: An unique ID identifying this Game in RAWG Database
  • slug: An unique slug identifying this Game in RAWG Database
  • name: Name of the game
  • metacritic: Rating of the game on Metacritic
  • released: The date the game was released
  • tba: To be announced state
  • updated: The date the game was last updated
  • website: Game Website
  • rating: Rating rated by RAWG user
  • rating_top: Maximum rating
  • playtime: Hours needed to complete the game
  • achievements_count: Number of achievements in game
  • ratings_count: Number of RAWG users who rated the game
  • suggestions_count: Number of RAWG users who suggested the game
  • game_series_count: Number of games in the series
  • reviews_count: Number of RAWG users who reviewed the game
  • platforms: Platforms game was released on. Separated by ||
  • developers: Game developers. Separated by ||
  • genres: Game genres. Separated by ||
  • publishers: Game publishers. Separated by ||
  • esrb_rating: ESRB ratings
  • added_status_yet: Number of RAWG users had the game as "Not played"
  • added_status_owned: Number of RAWG users had the game as "Owned"
  • added_status_beaten: Number of RAWG users had the game as "Completed"
  • added_status_toplay: Number of RAWG users had the game as "To play"
  • added_status_dropped: Number of RAWG users had the game as "Played but not beaten"
  • added_status_playing: Number of RAWG users had the game as "Playing"

Acknowledgements

Thanks to RAWG for providing easy to use and fast API
Icon made by Good Ware from www.flaticon.com

Inspiration

With this data, one can create a game recommendation platform as well as drawing insights about the gaming industry and gaming trends.