Time series analysis and prediction of competitive level play and card prices for Magic: The Gathering.

This project aims to provide a comprehensive database combining tournament level metagame data, card prices, and card info (CMC, printed sets, types, etc.). In addition to a series of investigations carried out on said data motivated by the following questions:

• How have the prices for cards and decks in non-rotating formats changed over time?
• How have format metagame compositions changed over time? Can they be predicted using current metagame compositions?
• How does a card seeing more or less play overtime impact its price?
• What impact do reprints have on the price of MTG cards? Can this impact be predicted?

This project was build using Python 3.8 and the libraries listed under requirements.txt. They can be installed with pip install -r requirements.txt.

The source root directory this project was built on is the parent directory where all top level files and subdirectories, such as the README.md reside. Any scripts that are run should be run as modules formatted like python -m path.to.script insert_arguments_here (dots instead of back slashes, drop the .py extension).

The size of the database is larger than the free tier of GitHub's Large File Storage, so to acquire a copy of the for your own purposes - the following must be done:

1. Setup Python environment as described above
2. Install Postgres 11
3. One of the following:
   1. Request a dump of the database from myself via Github or one of my other contact points. Load the dump into the database with database/db_init.py. This creates a new database with data intact.
   2. Initialize a blank copy of the database under database/db_init.py, after having installed Postgres. Repopulate the database by rerunning the scrapping scripts under scrapping/scrap_all.py.

Fair warning, attempting to repopulate the database manually may take upwards of 1-2 days, depending on the capabilities of your system. Recommended you adjust the multiprocessing parameters of each scrapper, the more the faster scrapping will go.

The configuration info used for the database is under database/db-config.yaml. Read database/db_init.py for instructions on how to run the script

This directory contains all the scripts and other supporting files necessary for setting up a blank version of the database, backing up an existing version of the database, and loading login info. More information can be found in the documentation of each file under this subdirectory.

The directory contains all the scripts necessary for rebuilding or updating the database by scrapping and parsing the appropriate sites, and saving their data in the database. More info can be found in the documentation of each file under this subdirectory.

This directory contains all the scripts necessary for recreating the analyses attempting to answer the motivating questions of this project. More info can be found in the documentation of each file under this subdirectory, as well as in the process reports for each analysis underlying it's assumptions and usage.

This project's code and data are completely free to use, modify, copy, or redistribute under the GPL3 license. However due to the size of the database a copy is not hosted on Github, but copies are available per request.

• Data storage and interfacing provided by Postgres 11.5 and psycopg2 2.3.8.
• Web scrapping and API calls with Requests and BeautifulSoup.
• Data visualization provided by Matplotlib.
• Machine learning and related utilities provided by Sci-Kit Learn and SciPy

• Tournament data from mtgtop8.
• Pricing data from MTGGoldfish
• Card info from the Scryfall API.