Scrap market historical prices for csgo items.
To launch :
-fill "credentials.json" with a steam account
-exectute "processing.py" (ex : "nohup python3 processing.py &")
The project returns these files:
-"rawSkinData.pickle" : every items of the csgo steam market
-"skinData.pickle" : same after processing (filter some items, remove duplicates)
-"rawHistoryData.pickle" : raw historical values for the items in the skinData
-"historyData.pickle" : formatted historical value, at daily frequency (when many values on a same day, we sum sold quantities and takes the weighted mean of the medians prices)
Performance is an issue on group_by operation. It works well with 8Gbytes of ram
Using python 3.8.10