lenapheno / ds-utils Goto Github PK

Python utility functions for data science.

• Base
  • R-like table function
• Explore
  • Missing value percentage of all columns in df
  • Get all unique keys in df
  • Count * group by columns
  • Customized describe() for numerical columns
  • Describe a numerical column given the value of a categorical column
  • Describe categorical columns including distribution of values
  • Describe a categorical column given the value of a categorical column
• Stats Vector-based and dataframe-based stat functions
  • Correlation matrix
  • Entropy and mutual information
  • t-test
  • Chi-square test
  • ANOVA
• Math
  • Log and inverse-log of different bases
  • Median absolute deviation (MAD)
• Preprocessing
  • Cast type for mulitple columns in bulk
  • Fill-in missing values in bulk
  • One-hot encoding and label encoding for training and test
• Evaluation
  • Sklearn scorer for log-scale target variable
• Visualization Dataframe oriented plot functions
  • histogram
  • scatterplot
  • barplot of distribution of categorical columns
  • boxplot of a numerical column given values of a categorical columns
  • stackedbarplot of a categorical column given values of another categorical column
  • pairplot of all columns
  • plot all rows of dataframe
  • plot explained variance ratio in PCA

Back to Home Page