generateData.py

# GENERATING THE DATA FOR CSV FILE
import numpy as np
from random import random
import csv



# Generates a list containing n-number of factors
def populate_factors(n):
	s = 'Factor '
	f = [s+str(i) for i in range(1,n+1)]
	return f

# Randomly generates a list of sigma values (std deviations) corresponding to the number of factors between [0.1,0.3] 
def get_std_devs(factors):
	std_devs = []
	for i in factors:
		# Sampling from adding a normal and a uniform distribution to get the standard deviations for each factor
		v = np.random.normal(0,.05,1) + np.random.uniform(0.1,0.25,1)
		std_devs.append(v)
	return std_devs


# Samples from a normal distribution n-times to get the pct change for n-number of data points using the sigma values generated
def get_pct_changes(n,std_devs):
	p = []
	for dev in std_devs:
		# s is a list, so pct changes is a list of lists
		s = np.random.normal(0,dev,n)
		p.append(s)
	return p


# Calculates the returns on index based on time-units
def get_returns(pct_changes,numDataPoints):
	rets = []
	
	num_returns = numDataPoints
	i = 0
	tmp = 0
	while(i!=num_returns):
		summing = 0
		for factor_change in pct_changes:
			# MUST FIGURE OUT CORRECT RANGES FOR THE WEIGHTS AND THEIR CORRESPONDING FACTORS
			tmp = i/100
			w = np.random.uniform(0.1+tmp,1.4+tmp)
			summing+=w*factor_change[i]
			# print(factor_change[i])
		rets.append(summing)
		i+=1
	return rets



def write_csv():
	myData = [factors,pct_changes,returns]
	myFile = open('testingScript.csv', 'w')
	with myFile:
		writer = csv.writer(myFile)
		writer.writerows(myData)


numFactors=100
numDataPoints = 10000


factors = populate_factors(numFactors)
std_devs = get_std_devs(factors)
pct_changes = get_pct_changes(numDataPoints,std_devs)

returns = get_returns(pct_changes,numDataPoints)
print(max(returns))
print(min(returns))
print(sum(returns)/len(returns))
write_csv()