Codecademy export

script.py

import codecademylib
import pandas as pd

visits = pd.read_csv('visits.csv',
                     parse_dates=[1])
cart = pd.read_csv('cart.csv',
                   parse_dates=[1])
checkout = pd.read_csv('checkout.csv',
                       parse_dates=[1])
purchase = pd.read_csv('purchase.csv',
                       parse_dates=[1])
#ppl visit + cart or not
visit_to_cart = pd.merge(visits, cart, how = 'left')
num_visit = len(visit_to_cart)
print(num_visit)
#num of ppl visiting :2052
visit_to_cart['no_cart'] = ~ visit_to_cart.cart_time.isnull()
num_no_cart = visit_to_cart.no_cart.sum()
print(num_no_cart)
#num of ppl no_cart:1652
percent_visit_no_cart = float(num_no_cart)/num_visit
print(percent_visit_no_cart)
#0.805068226121

cart_to_checkout = cart.merge(checkout, how = 'left')
num_cart = len(cart_to_checkout)
print(num_cart)
#num of people cart :602 (some people put it in the cart without visiting first???)
no_checkout = cart_to_checkout[cart_to_checkout.checkout_time.isnull()]
num_no_checkout = len(no_checkout)
print(num_no_checkout)
#num of ppl no_checkout :126
percent_cart_no_checkout = float(num_no_checkout)/num_cart
print(percent_cart_no_checkout)
#0.209302325581

all_data = visits.merge(cart, how = 'left').merge(checkout, how = 'left').merge(purchase, how = 'left')
print(len(visits))