In these notebooks PyMC3 is used to model the number of registered KeyForge decks. Data on the weekly registragions was provided by Duk from Archon Arcana. Names of Dark Tidings decks for the second part to estimate the total number of decks printed were provided by Saluk, another member of Archon Arcana
For more details read the corresponding blog post on 4DCu.be here
Using PyMC3 we model the number of registrations including releasing new sets and the global outbreak of COVID-19. Once a good model was found, we also check how many decks would have been registered if COVID-19 didn't put the world in lockdown.
You can open the notebooks on Binder.
To get the code running locally the easiest option is using Anaconda
git clone https://github.com/4dcu-be/BayesianSalesAnalysis
cd BayesianSalesAnalysis
conda env create -f environment.yml
conda activate pymc3
jupyter notebookIn case you are running this on windows you might need to install libpython in the environment and you will need Visual Studio Code 2017 build tools (with the C tools, check the options).