This repository holds a feature generator for financial time series data. Given a dataframe with datetime index, data is resampled to given frequencies. Customize feature periods in featureConfig.py.
Instantiate a FeatureGenerator object:
from .featureGenerator import FeatureGenerator
import featureConfig
fg = FeatureGenerator(featureConfig)
Pass in your dataframe:
df = pd.read_csv('example_data.csv')
fg.calculate_all_features(df)
Your dataframe now contains all desired features as new columns.
Additionally, you may access specific feature generation functions in a static context:
import .featureGenerator as featgen
df = pd.read_csv('example_data.csv')
periods = [10, 20, 30]
featgen.add_sma(df, 'close', periods)
df now contains 3 new columns, close_sma_10, close_sma_20, and close_sma_30.