In [205]:

    

%matplotlib inline
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns


    

In [206]:

    

# defining my simulation constants
paycheck_norm = 4200
paycheck_std = 1000


    

In [207]:

    

# a regular transaction, for example a paycheck, electric bill or coffee
class RegTrans(object):
    def __init__(self, norm=-20, std=-2, period='weekly', date='random', time='random', desc='coffee'):
        self.amount = round(np.random.normal(loc=norm, scale=std), 2)
        self.period = period
        self.desc = desc
        if date == 'random':
            if self.period == 'monthly':
                self.date = np.random.randint(1,28)
            if self.period == 'weekly':
                self.date = np.random.randint(1,7)


    

In [208]:

    

# paycheck as a regular transaction
class Paycheck(RegTrans):
    def __init__(self):
        RegTrans.__init__(self, norm=paycheck_norm, std=paycheck_std, period='monthly', date='random', desc='paycheck')


    

In [218]:

    

# elec bill as a regular transaction
class ElecBill(RegTrans):
    def __init__(self):
        RegTrans.__init__(self, norm=-100, std=10, period='monthly', date='random', desc='elec_bill')


    

In [219]:

    

# a user, with a paycheck, bills and other regular payments
class User(object):
    def __init__(self):
        self.userid = np.random.randint(10**10) 
        self.paycheck = Paycheck()
        self.elec_bill = ElecBill()
        self.reg_trans = [self.paycheck, self.elec_bill]


    

In [220]:

    

# add a user's transactions to the DataFrame df
def add_user(df):
    user = User()
    for year in xrange(2013, 2016):
        for month in xrange(1, 12):
            for reg_trans in user.reg_trans:
                if reg_trans.period == 'monthly':
                    date = pd.datetime(year=year, month=month, day=reg_trans.date)
                    df.loc[len(df)] = [user.userid, reg_trans.amount, date, reg_trans.desc]


    

In [221]:

    

# define an empty DataFrame
trans = pd.DataFrame(columns=['userid', 'amount', 'date', 'desc'])


    

In [224]:

    

# add 50 users to the DataFrame
for i in xrange(50):
    add_user(trans)


    

In [225]:

    

len(trans)


    

    
Out[225]:





9900

In [226]:

    

trans.head()


    

    
Out[226]:






  

    

      

      
userid
      
amount
      
date
      
desc
    
  
  

    

      
0
      
4140899731
      
4774.22
      
2013-01-12
      
paycheck
    
    

      
1
      
4140899731
      
-122.77
      
2013-01-25
      
elec_bill
    
    

      
2
      
4140899731
      
4774.22
      
2013-02-12
      
paycheck
    
    

      
3
      
4140899731
      
-122.77
      
2013-02-25
      
elec_bill
    
    

      
4
      
4140899731
      
4774.22
      
2013-03-12
      
paycheck
    
  

In [229]:

    

paychecks = trans[trans.desc == 'elec_bill'].groupby(trans.userid).mean().amount
paychecks.hist(bins=20)


    

    
Out[229]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f46af1e6850>






    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]: