Gran Turismo Scrape

Utilities to scrape and analyze data about Gran Turismo 2.

My intent was to visualize some of the things I noticed in the game:

• large skew towards Japanese cars (both in terms of total cars, and prize cars)
• large skew towards 90's cars

Both of these properties are not to be unexpected coming from a Japanese game development company in 99, yet I still wanted to show it through data.

Peek at data

Table with all car information available in GT2 Wiki

Plot showing the distribution of all cars, and also prize cars, per manufacturing country (see gt/clean.py to see how to clean the information to be able to plot this):

Download data

Currently hosting scraped data on Dropbox:

• 2020-09-23 V2: dropbox download
  • added clean table "cars-clean.csv"
  • cleaned the following fields:
    • str to int: height, width, length, displacement, max power, power weight ratio, weight
    • str to str: drivetrain
  • added new fields:
    • "Country" and "Continent" of origin
    • "IsPrizeCar": is the car a prize car
    • "IsRaceModifiable": is the car race modifiable
    • "Date": manufacturing date
• 2020-09-12 V1: dropbox download

Installation instructions

Assuming you use conda, and create a conda environment called gt:

conda create --name gt python=3.6
conda activate gt
conda install -c conda-forge --file requirements.txt

Notes

• I used jupytext to convert my jupyter notebook to .py file, this is very convenient as it is way more git friendly / allows much easier version control.
• I learned scraping from the following blog post introducing to beautiful soup.