Retail Projects 1 - Laptop Sales Analysis Intro 1 - Section 117

Search the site...

Sentry Page Protection

Geo-targeting Project [17-25]

Let's look at an overall picture of how far customers travel to buy computers.

proc univariate data=sales_q1_loc_pos4;
var distance;
histogram;
run;

Did the code fail?

Copy and run the code below to create the input data sets.

proc import datafile = "/folders/myfolders/Geo-targeting/Data Files/Sales_Q1.csv"
 out = Sales_Q1
 dbms = CSV
 replace
 ;
run;

proc import datafile = "/folders/myfolders/Geo-targeting/Data Files/locations.csv"
 out = locations
 dbms = CSV
 replace
 ;
run;

proc import datafile = "/folders/myfolders/Geo-targeting/Data Files/postcode.csv"
 out = postcode
 dbms = CSV
 replace
 ;
run;

proc sql;
create table sales_q1_loc as
select a.*, b.os_x as store_x, b.os_y as store_y
from sales_q1 a, locations b
where a.store_postcode = b.postcode;
quit;

proc sql;
create table sales_q1_loc_pos as
select a.*, b.os_x as cust_x, b.os_y as cust_y
from sales_q1_loc a left join postcode b
on a.customer_postcode = b.postcode;
quit;

filename postcode '/folders/myfolders/Geo-targeting/Data Files/open_postcode_geo.csv';

data all_postcode (keep=postcode os_x os_y);
infile postcode dsd firstobs=2;
input postcode : $8.
      col1 : $10.
      col2 : $5.
      os_x os_y
      col3 
      area : $8.;
if area = "England" and os_x^=.;
run;

proc sql;
create table sales_q1_loc_pos2 as
select a.*, b.os_x as cust_x1, b.os_y as cust_y1
from sales_q1_loc_pos a left join all_postcode b
on a.customer_postcode = b.postcode;
quit;

data sales_q1_loc_pos3;
set sales_q1_loc_pos2;

if cust_x = . then cust_x = cust_x1;
if cust_y = . then cust_y = cust_y1;

drop cust_x1 cust_y1;
run;

The mean distance traveled to make purchases at our store is 3678 meters (3.6 km)

In the histogram, you will see that there is quite a significant drop off in distance at 4.5 km:

We will use this as a reference point. Our direct marketing campaign will focus on the areas that are within 4.5 km of our stores.