ds_functions_package

package for storing useful data science functions

pip install git+https://github.com/taylor-m/ds_functions_pkg

my_functions_pkg.supervised_regression.

model_cv_stats

(X, y, test_size, random_state, alphas, cv)

OLS, RidgeCV, LassoCV, ElasticNetCV
prints: best alpha, train R^2, test R^2, MAE, MSE, RMSE, MAPE
returns: df of the cv stats

predictions_df

(X_test, y_test, y_preds)

returns predictions df and plotly express fig object
data frame = X_test.copy() + y_true + y_preds + residuals + abs_residuals
descending order by abs_residuals to view highest model mistakes made & index

plot_sced_test

(df, target, features)

checks homoscedasticity assumption with plots

top_corrs

(df, column, n=10)

returns top (n) absolute correlated feature vars vs. target var

plot_top_corrs

(df, column, n=10)

plots the topp_corrs() using a facet grid

tukey_outliers

(df, var)

prints outliers at each threshold (1-5)

print_vif

(feature_df)

check multicollinearity assumption; prints VIF for each feature in feature_df

my_functions_pkg.stats.

get_95_ci

(x1, x2)

95% confidence interval

cles_ind

(x1, x2)

common language effect size

calc_non_param_ci

(x1, x2, alpha=0.05)

2 median group test confidence interval