coinmetrics_features_prep.py

x_cols = ['AdrActCnt', 'BlkCnt', 'BlkSizeByte', 'BlkSizeMeanByte', 'CapMVRVCur',
          'CapMrktCurUSD', 'CapRealUSD', 'DiffMean', 'FeeMeanNtv', 'FeeMeanUSD', 'IssContNtv',
          'IssContPctAnn', 'IssContUSD', 'IssTotNtv', 'IssTotUSD', 'NVTAdj',
          'NVTAdj90','TxCnt', 'TxTfrCnt', 'TxTfrValAdjNtv', 'TxTfrValAdjUSD',
          'TxTfrValMeanNtv', 'TxTfrValMeanUSD', 'TxTfrValMedNtv',
          'TxTfrValMedUSD', 'TxTfrValNtv', 'TxTfrValUSD', 'VtyDayRet180d',
          'VtyDayRet30d', 'VtyDayRet60d']

y_col = ['PriceUSD']

xvars_df = btc_df[x_cols]
xvars_df.replace([np.inf, -np.inf], np.nan)
xvars_df.dropna(inplace=True)

xvars_df = (xvars_df - xvars_df.rolling(60).mean())/xvars_df.rolling(60).std()  # normalize xvars by zscore
xvars_df = xvars_df.shift(1)  # lag x vars by 1
btc_px = btc_df[y_col]
btc_px.columns = ['BTC_PX']
y = btc_px.pct_change(1)

y.replace([np.inf, -np.inf], np.nan)
y.dropna(inplace=True)

y.columns = ['Y_VAR']
reg_dat = pd.concat([xvars_df,y],axis=1).dropna()