python.py

import numpy as np
#import seaborn as sb
import matplotlib.pyplot as plt
import statistics as st
import pandas as pd
import scipy as sp
from scipy.stats import norm
import random

def norm_converge(samp_siz, num_sims, exp = True, param = 0.3):
  
    mean_vec = np.empty([num_sims])
  
    for i in range(num_sims):
  
      if exp == True:
          samp = np.random.exponential(scale = param, size = samp_siz)
      else:
          samp = np.random.weibull(a = param, size = samp_siz)
      
      mean_vec[i] = st.mean(samp)
  
    return(mean_vec)

# Define function sampler

def sampler(df, colname,N, alpha, mu_0):

  rows = random.sample(range(df.shape[0]), N)
  samp = df.iloc[rows]
  
  col = samp[colname]
  
  my_mean = (1/len(col)) * sum(col)
  my_var = 1/(len(col) - 1) * np.sum((col - np.mean(col))**2)
  my_std = np.sqrt(my_var)
  
  # Find z value at alpha / 2
  z = norm.ppf(1-alpha / 2, loc = 0, scale = 1)
  
  mu_l, mu_h = my_mean - z * my_std / np.sqrt(len(col)), my_mean + z * my_std / np.sqrt(len(col))
  
  W = (my_mean - mu_0) / (my_std) / np.sqrt(len(col))
  
  Test = abs(W) <= z
  
  out = dict({"mean": my_mean, 
              "var": my_var,
              "std_dev":my_std,
              "mu_low": mu_l,
              "mu_high": mu_h,
              "Test stat": W,
              "Reject H_O": Test})

  return(out)