linear_regression_analysis.py

import pandas as pd
import statsmodels.api as sm

def analysis_lr(df:pd.DataFrame, l_columns_X:list, column_Y:str, alpha:float = .01)->pd.DataFrame:
  # validate arguments
  assert isinstance(df, pd.DataFrame) and len(df) > 0
  for c in l_columns_X + [column_Y]:
    assert c in df.columns.tolist()
  # data preparation
  X = df[l_columns_X]
  X = sm.add_constant(X)  # Agrega la constante
  Y = df[column_Y]
  # fit
  modelo = sm.OLS(Y, X).fit()
  # collect analysis results
  coeficientes = modelo.params
  std_err = modelo.bse
  p_values = modelo.pvalues
  conf_int = modelo.conf_int()    
  r_squared = modelo.rsquared
  # storing results
  conf_int.columns = ["ci_i", "ci_f"]
  dfr = pd.concat([
      pd.DataFrame(coeficientes, columns = ["coefs"]),
      pd.DataFrame(std_err, columns = ["std_err"]),
      pd.DataFrame(p_values, columns = ["p_values"]),
      conf_int
  ], axis = 1)
  dfr["r_squared"] = np.full(len(dfr), r_squared)
  dfr["is_reliable"] = dfr["p_values"] < alpha
  #return
  return dfr