MARKETING &

RETAIL ANALYTICS
z

SANDYA VB
22-08-2021
z
PROBLEM STATEMENT

▪ An automobile parts manufacturing company has collected data

of transactions for 3 years. They do not have any in-house data
science team, thus they have hired you as their consultant. Your
job is to use your magical data science skills to provide them
with suitable insights about their data and their customers.

▪ DATA: Sales_Data.xlsx
 z
DATA DICTIONARY

CUSTOMERNAM
ORDERNUMBER : Order Number customer
E:
QUANTITYORDERED
Quantity ordered PHONE : Phone of the customer
:
PRICEEACH : Price of Each item ADDRESSLINE1 : Address of customer
ORDERLINENUMBER
order line CITY : City of customer
:
SALES : Sales amount POSTALCODE : Postal Code of customer
ORDERDATE : Order Date COUNTRY : Country customer
DAYS_SINCE_LASTO CONTACTLASTN
Days_ Since_Lastorder Contact person customer
RDER : AME :
CONTACTFIRST
STATUS : Status of order like Shipped or not Contact person customer
NAME :
Size of the deal based on Quantity and
PRODUCTLINE : Product line – CATEGORY DEALSIZE :
Item Price
Manufacturer's Suggested
MSRP :
Retail Price
PRODUCTCODE : Code of Product
z
READING THE DATASET
The dataset is read using the read function.
z
EXPLORATORY ANALYSIS AND
INFERENCES

The dataset is measured using central measures for all the columns with integer
values.

It tells how the data is been distributed, deviated or centrally aligned.

 z
▪ This gives the info of all the
columns of the dataset.

▪ Most of the columns are of object

type, and the rest are of int and
float type.

▪ Column
DAYS_SINCE_LASTORDER is
dropped.

▪ The datatype of ORDERDATE is

changed to datetime64[ns]
format.
 z

▪ Df.shape gives the shape of the

dataset that is it gives the total
number of rows and columns in
the dataset.

▪ The dataset has 2747 rows and

19 columns, after dropping
DAYS_SINCE_LASTORDER.

▪ Otherwise the original dataset

had 2747 rows and 20 olumns.
 z

▪ Here we see that the dataset

does not have any null values.

▪ If there was any missing values

present or any duplicate values,
we would have treated it before
performing any calculations.
z
UNIVARIATE ANALYSIS

▪ The integer columns of the dataset are represented in graphically

representation.

▪ Few of the columns have outliers.

▪ Most of the data is normally distributed.

 z
BIVARIATE ANALYSIS

▪ For Bi-variate analysis we

perform pairwise relationship.
 z
MULTIVARIATE ANALYSIS

▪ For Multivariate analysis we

perform heatmap for the
different columns.

▪ PRICEEACH is highly correlated

with MSRP.

▪ SALES is highly correlated to

PRICEEACH
 z
TRENDS IN SALE

▪ Sales is highest in the 4Th

Quarter for the year 2018 and for
the year 2019 when compared to
all other quarters.

▪ For the year 2020 sales is highest

in the 2nd Quarter.

▪ For the rest of the quarters, sales

is low and is on an increasing
and decreasing trend.
 z

▪ We see that on day basis,

Sunday has the highest sales
followed by Tuesday.

▪ But Thursday has the least sales

compared to other week days.
 z
SALES ACROSS DIFFERENT CATEGORIES OF
THE GIVEN DATA

▪ We see that in sales across the

country, USA has the maximum
sales followed by Spain.

▪ The county with least or lowest

sales is Ireland.
 z

▪ Most of the sales is happening in medium deal size. This means that most of the
orders are of medium size and not too large or small.

▪ The flow of sales as per deal size is Medium > Small > Large.

▪ This means that least orders are of large deal size.

▪z From the chart below we see that most of the sales is
happening in Classic Cars.

▪ The flow of sales as per product line is Classic Cars > Vintage
cars> Trucks and buses> motorcycles> planes> ships> trains.

▪ This means that least orders are of trains and highest of classic
cars
▪ From the below chart we see that most of the orders are shipped.
z

▪ Few orders have been cancelled where as few are in process.

▪ But we also see that orders which are under dispute are been resolved
but still some are yet to be resolved.
z
INFERENCE

▪ From the chart trends across sales we come to know that the sales are maximum during the
4th quarter, to increase over all sales across all the quarters offers or discounts can be given
to the customers.

▪ To increase the sales in countries with least sales, mega offers sales with low EMI facilities
can be given to promote the sales.

▪ The large size deals are lowest and almost stagnant. Steps should be taken to promote and
attract the customers to buy more of large size deals.

▪ Classic cars have majority sales whereas trucks and buses sales can be expanded.

▪ Should focus on cancelled orders to check why it was cancelled and work on it. Pending
disputed orders should be resolved at the earliest so it doesn’t gives a negative feel to the
customer.

▪ Focus should be on expanding sales by taking in more customers by giving more offers like
low EMI, festival sale, mega sale, etc.
z
 z

▪ Customer Segmentation Using

RFM Analysis .
▪ Using CRM, we assigned weights of 0.33 to each of the factors.
z
▪ Post this, we fitted the data for the weights and obtained the scores.
z
BEST CUSTOMERS
z
CUSTOMERS WHO ARE ON THE VERGE OF
CHURNING
z
LOST CUSTOMERS
z
LOYAL CUSTOMERS
z

TABLEAU LINK:
▪ RFMProject-
MIlestone1://public.tableau.com/authoring/MRAprojectMilestone
1/trendsacrosssales#1

▪ https://public.tableau.com/authoring/MRAprojectMilestone1/tren
dsacrosssales#1