San Diego Burrito Analytics: California burritos

Scott Cole

27 August 2016

This notebook formats the burrito data set in order to analyze California burritos
The data set is format so that it can be read into Tableau to make a visualization
Analyze the different kinds of California burritos served at the different locations (with guac?)

Default imports



In [1]:

    
%config InlineBackend.figure_format = 'retina'
%matplotlib inline

import numpy as np
import scipy as sp
import matplotlib.pyplot as plt
import pandas as pd

import seaborn as sns
sns.set_style("white")

Load data



In [2]:

    
import util2
df, dfRestaurants, dfIngredients = util2.load_burritos()
N = df.shape[0]

Limit data to California burritos



In [3]:

    
dfCali = df.loc[df.Burrito.str.contains('.*[Cc]ali.*')]
dfCaliIngredients = dfIngredients.loc[df.Burrito.str.contains('.*[Cc]ali.*')][['Beef','Pico','Guac','Cheese','Fries','Sour cream','Chicken']]
dfRestaurants=dfRestaurants.reset_index().drop('index',axis=1)
dfCaliRestaurants = dfRestaurants.loc[[i for i,x in enumerate(dfRestaurants.Location) if x in dfCali.Location.unique()]]

Process Cali burrito data: Averages for each restaurant



In [4]:

    
dfCaliAvg = dfCali.groupby('Location').agg({'Cost': np.mean,'Volume': np.mean,'Hunger': np.mean,
                                'Tortilla': np.mean,'Temp': np.mean,'Meat': np.mean,
                                'Fillings': np.mean,'Meat:filling': np.mean,'Uniformity': np.mean,
                                'Salsa': np.mean,'Synergy': np.mean,'Wrap': np.mean,
                                'overall': np.mean, 'Location':np.size})
dfCaliAvg.rename(columns={'Location': 'N'}, inplace=True)
dfCaliAvg['Location'] = list(dfCaliAvg.index)



In [5]:

    
# Calculate latitutude and longitude for each city
import geocoder
addresses = dfCaliRestaurants['Address'] + ', San Diego, CA'
lats = np.zeros(len(addresses))
longs = np.zeros(len(addresses))
for i, address in enumerate(addresses):
    g = geocoder.google(address)
    Ntries = 1
    while g.latlng ==[]:
        g = geocoder.google(address)
        Ntries +=1
        print 'try again: ' + address
        if Ntries >= 5:
            raise ValueError('Address not found: ' + address)
    lats[i], longs[i] = g.latlng
    
# # Check for nonsense lats and longs
if sum(np.logical_or(lats>34,lats<32)):
    raise ValueError('Address not in san diego')
if sum(np.logical_or(longs<-118,longs>-117)):
    raise ValueError('Address not in san diego')









    



try again: 377 Carlsbad Village Dr, San Diego, CA
try again: 377 Carlsbad Village Dr, San Diego, CA
try again: 377 Carlsbad Village Dr, San Diego, CA
try again: 820 Birmingham Dr, San Diego, CA
try again: 820 Birmingham Dr, San Diego, CA



In [6]:

    
# Incorporate lats and longs into restaurants data
dfCaliRestaurants['Latitude'] = lats
dfCaliRestaurants['Longitude'] = longs
# Merge restaurant data with burrito data
dfCaliTableau = pd.merge(dfCaliRestaurants,dfCaliAvg,on='Location')



In [7]:

    
dfCaliTableau.to_csv('cali_now.csv')