Scribd
Upload
358 views

Project 4: Retail Analysis With Walmart Data

The document outlines an analysis of Walmart sales data with the following objectives: 1. Analyze basic statistics like maximum sales by store, stores with highest variability in sales, and growth in Q3 2012. 2. Identify holidays with above average sales compared to non-holidays. 3. Build linear regression and logistic regression models to forecast sales for Store 1 and examine the impact of variables like date, CPI, unemployment, and fuel price.

Uploaded by

bhadav
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
358 views

Project 4: Retail Analysis With Walmart Data

The document outlines an analysis of Walmart sales data with the following objectives: 1. Analyze basic statistics like maximum sales by store, stores with highest variability in sales, and growth in Q3 2012. 2. Identify holidays with above average sales compared to non-holidays. 3. Build linear regression and logistic regression models to forecast sales for Store 1 and examine the impact of variables like date, CPI, unemployment, and fuel price.

Uploaded by

bhadav
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 12

Project 4  Retail Analysis with Walmart Data

Objective:

Basic Statistics tasks


 Which store has maximum sales.
 Which store has maximum standard deviation i.e., the sales vary a lot. Also, find out the
coefficient of mean to standard deviation
 Which store/s has good quarterly growth rate in Q3’2012.
 Some holidays have a negative impact on sales. Find out holidays which have higher
sales than the mean sales in non-holiday season for all stores together.
 Provide a monthly and semester view of sales in units and give insights.

Statistical Model
For Store 1 – Build  prediction models to forecast demand
 Linear Regression – Utilize variables like date and restructure dates as 1 for 5 Feb 2010
(starting from the earliest date in order). Hypothesize if CPI, unemployment, and fuel price have
any impact on sales.
 Change dates into days by creating new variable.
Select the model which gives best accuracy.
CODE {}

#Import Required Libraries

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
from patsy import dmatrices
import sklearn
import seaborn as sns

import os
print(os.getcwd())

walmartdata = pd.read_csv('Walmart_Store_sales.csv')
walmartdata.head()

walmartdata_groupby = walmartdata.groupby('Store')['Weekly_Sales'].sum()

# Which store has Maximum sales and Total sales.

print("Store Number {} has maximum Sales. Sum of Total Sales


{}".format(walmartdata_groupby.idxmax(),walmartdata_groupby.max()))
#Which store has maximum standard deviation i.e., the sales vary a lot. Also, find out the
coefficient of mean to standard deviation.

walmartdata_std = walmartdata.groupby('Store').agg({'Weekly_Sales':'std'})

print("Store Number {} has maximum Standard Deviation. STD


{}".format(walmartdata_std['Weekly_Sales'].idxmax(),walmartdata_std['Weekly_Sales'].max()))
walmartdata_std = walmartdata.groupby('Store').agg({'Weekly_Sales':['mean','std']})

walmartdata_std.head()

walmartdata_Q32012 = walmartdata[(pd.to_datetime(walmartdata['Date']) >= pd.to_datetime('07-


01-2012')) & (pd.to_datetime(walmartdata['Date']) <= pd.to_datetime('09-30-2012'))]

walmartdata_growth = walmartdata_Q32012.groupby(['Store'])['Weekly_Sales'].sum()

print("Store Number {} has Good Quartely Growth in Q3'2012


{}".format(walmartdata_growth.idxmax(),walmartdata_growth.max()))
#To Find out holidays which have higher sales than the mean sales in non-holiday season for all
stores together

stores_holiday_sales = walmartdata[walmartdata['Holiday_Flag'] == 1]

stores_nonholiday_sales = walmartdata[walmartdata['Holiday_Flag'] == 0]

#Stores Sales in Super Bowl Day


#Super Bowl: 12-Feb-10, 11-Feb-11, 10-Feb-12, 8-Feb-13

stores_holiday_sales_superBowl =
stores_holiday_sales[(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('12-02-
2010')) |(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('11-02-2011'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('10-02-2012'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('08-02-2013'))]
#Stores Sales in Labour Day
#Labour Day: 10-Sep-10, 9-Sep-11, 7-Sep-12, 6-Sep-13

stores_holiday_sales_labourDay =
stores_holiday_sales[(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('10-09-
2010')) |(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('09-09-2011'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('07-09-2012'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('06-09-2013'))]

#Stores Sales in Thanks Giving


#Thanksgiving: 26-Nov-10, 25-Nov-11, 23-Nov-12, 29-Nov-13

stores_holiday_sales_thanksgiving =
stores_holiday_sales[(pd.to_datetime(stores_holiday_sales['Date]) == pd.to_datetime('26-11-
2010')) |(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('25-11-2011'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('23-11-2012'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('29-11-2013'))]
#Stores Sales in Christmas
# Christmas: 31-Dec-10, 30-Dec-11, 28-Dec-12, 27-Dec-13

stores_holiday_sales_Christmas =
stores_holiday_sales[(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('31-12-
2010')) |(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('30-12-2011'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('28-12-2012'))|
(pd.to_datetime(stores_holiday_sales['Date']) == pd.to_datetime('27-12-2013'))]

stores_nonholiday_sales_mean =
stores_nonholiday_sales.groupby(['Date']).agg({'Weekly_Sales':'mean'}).reset_index()
stores_holiday_sales_sum =
stores_holiday_sales.groupby(['Date']).agg({'Weekly_Sales':'sum'}).reset_index()
for row in stores_holiday_sales_sum.itertuples():
for row1 in stores_nonholiday_sales_mean.itertuples():
if row.Weekly_Sales > row1.Weekly_Sales:
print("On this Date {} Holiday Sales is greater than Non Holiday Sales and the Sales :-
{}".format(row.Date,row.Weekly_Sales))
break;
print("Super Bowl Day Sale",stores_holiday_sales_superBowl['Weekly_Sales'].sum())
print("Labour Day Sale",stores_holiday_sales_labourDay['Weekly_Sales'].sum())
print("Thanksgiving Day Sale",stores_holiday_sales_thanksgiving['Weekly_Sales'].sum())
print("Christmas Day Sale",stores_holiday_sales_Christmas['Weekly_Sales'].sum())

x_features_object = walmartdata[walmartdata['Store'] ==1][['Store','Date']]


date_obj = walmartdata[walmartdata['Store'] ==1][['Date']]
date_obj.index +=1
x_features_object.Date = date_obj.index
x_features_object.head()
y_target = walmartdata[walmartdata['Store'] ==1]['Weekly_Sales']
y_target.head()

from sklearn.model_selection import train_test_split


x_train,x_test,y_train,y_test = train_test_split(x_features_object,y_target,random_state=1)

#Linear Regression – Utilize variables like date and restructure dates as 1 for 5 Feb #2010 (starting
from the earliest date in order).
#Hypothesize if CPI, unemployment, and #fuel price have any impact on sales.

from sklearn.linear_model import LinearRegression


linreg = LinearRegression()
linreg.fit(x_train,y_train)
feature_dataset = walmartdata[walmartdata['Store'] ==1][['Store','CPI','Unemployment','Fuel_Price'
]]
feature_dataset.head()

response_set_cpi = walmartdata[walmartdata['Store'] ==1]['CPI'].astype('int64')


response_set_unemployment = walmartdata[walmartdata['Store']==1]
['Unemployment'].astype('int64')

from sklearn.model_selection import train_test_split


x_train_cpi,x_test_cpi,y_train_cpi,y_test_cpi =
train_test_split(feature_dataset,response_set_cpi,random_state=1)
x_train_unemp,x_test_unemp,y_train_unemp,y_test_unemp =
train_test_split(feature_dataset,response_set_unemployment,random_state=1)

from sklearn.linear_model import LogisticRegression


logreg = LogisticRegression(max_iter=10000)
logreg.fit(x_train_cpi,y_train_cpi)
y_pred = logreg.predict(x_test_cpi)

logreg.fit(x_train_unemp,y_train_unemp)
y_pred_unemp = logreg.predict(x_test_unemp)
from sklearn import metrics
print(metrics.accuracy_score(y_test_cpi,y_pred))
print(metrics.accuracy_score(y_test_unemp,y_pred_unemp))

print('cpi actual :', y_test_cpi.values[0:30])


print('cpi Predicted :', y_pred[0:30])
print('actual Unemployment :', y_test_unemp.values[0:30])
print('Predicted Unemployment :', y_pred_unemp[0:30])
walmartdata['Day'] = pd.to_datetime(walmartdata['Date']).dt.day_name()
walmartdata.head()

You might also like