DataScience Help

ds.py

# First we need a data structure of a datframe with 3 columns ( visitID, landing_page, and converted )
# Then we need to run a probability loop for 10,000 iterations with a 50/50 split getting old or new page, and their respective
# probabilities of converting

# Borrowed this because it worked well
# https://stackoverflow.com/questions/439115/random-decimal-in-python
import decimal
def gen_random_decimal(i,d):
    return decimal.Decimal('%d.%d' % (random.randint(0,i),random.randint(0,d)))

# First create the dataframe tol hold the data
simulated_df_nnew = pd.DataFrame(columns=['user_id', 'landing_page', 'converted'])

# Create a dictionary to store the results:
dict = {}

# Now loop to fill the dictionary
for i in range(nnew):
    # Get a random probability to see if the page converts
    random_conversion = gen_random_decimal(0,100000000)
    #print(simulated_df_nnew)
    # Set to new page:
    landing_page = "new_page"
    # Set default to did not covert:
    conversion = 0
    # Compare to determine conversion
    if random_conversion >= (1.0 - convert_rate_pnew):
        conversion = 1
    # Append the results to the dictionary
    dict[i] = {"user_id": i, "landing_page":landing_page, "conversion": conversion}


# Finally build the dataframe from the dictionary
simulated_df_nnew = pd.DataFrame.from_dict(dict, "index")