Scribd
Upload
218 views

10.1 Time Series Analysis Sales Forecast

Uploaded by

senjaliya tanvi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
218 views

10.1 Time Series Analysis Sales Forecast

Uploaded by

senjaliya tanvi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 7

Practical Analytics Chapter 10 | Exercise 1 | Edition 3

10.1
Time Series Analysis - Sales
Forecast
Nitin Kalé, University of Southern California
Nancy Jones, San Diego State University

OBJECTIVE
The objective of this exercise is to use time series analysis to forecast Global Bike U.S. sales and interpret the
reliability of the forecasting model.

ACTIVITIES
 Import and prepare data.
 Configure forecasting models.
 Analyze and interpret output from models.

SOFTWARE PREREQUISITES
Access to SAP Analytics Cloud with Predictive Scenarios

DATA SET
Data file titled GBSales_transactions.xlsx

Scenario
Nina and other managers at Global Bike are getting ready to develop strategic plans and budgets for
the upcoming year. Therefore, Nina is interested in forecasting sales revenue for at least one year
from the date for which data are available. She will use time series analysis for forecasting.

Nitin Kalé & Nancy Jones © 2021 1 of 7


Practical Analytics Chapter 10 | Exercise 1 | Edition 3

Time Series Analysis


Time series analysis is a technique that analysts use to (a) uncover any implicit structure (patterns
or trends) in the data and (b) model that structure to make forecasts. The assumption is that the
future, at least in the short term, will continue the structure of the past. This technique is useful
wherever forecasting values such as sales quantities, airline passenger volume, economic metrics,
and traffic volume is needed.

1. Acquire and Aggregate the Data


The data set given to Nina includes sales transactions from 2008 through 2020. To create a
forecast, we will want to consolidate/aggregate the details into one-month time periods. We can
do this easily in a private model in SAC.
1. Select Stories from the menu on the left side of the screen.
2. Select Create New Canvas.
3. Select Add data.
a. Select Data uploaded from a File  Source Source File.
(1) Find the Excel sheet GBSales_transactions.xlsx and open.
b. Import.
c. On the Data tab of the Story do the following:
(1) Change Year and Month to dimensions
(2) Change Year and Month details.
(i) Change Data Type to String.
(ii) Check to be sure Statistical Type is set to Continuous.
(3) Concatenate Month and Year as shown below using the transformation function. The
name of the column will default to Month_Year.

Figure 1: Concatenate Month and Year

d. On the Story view, create a Table.


(1) Include Revenue as the measure and Currency, and Month_Year in that order as
Rows. Your results should look similar to those in Figure 2.

Nitin Kalé & Nancy Jones © 2021 2 of 7


Practical Analytics Chapter 10 | Exercise 1 | Edition 3

Figure 2: Create the Table

4. Export the results of the table using the Export function under the table’s More Actions. The
parameters for Export function are shown in .

Nitin Kalé & Nancy Jones © 2021 3 of 7


Practical Analytics Chapter 10 | Exercise 1 | Edition 3

Figure 3: Export the Table Results

a. Open the .csv file that was downloaded to your computer. It should look similar to the
following figure.

Figure 4: The .csv File from SAC

(1) Move the heading “REVENUE” to row 2 and delete row 1.


(2) Save the file as an Excel spreadsheet. (You may also save as a .csv.) I named mine
GBSales_transactions_agreggated.xlsx.

2. Create the Aggregated Data to a Data Set


SAC Predictive Scenario uses data sets or data models to model predictive analyses. Data sets
are models that are created for “personal” use; that is, they are not deployed out to the
organization. Models are more formal data structures and are usually created to be used by
many. In this assignment, we will convert the aggregated Global Bike sales data into a “Data
Set”.
1. Select Datasets from the menu on the left side of the screen.
2. Select Create New From a CSV or Excel File.
a. Select Source File.
b. Select GBSales_transactions_aggregated.xlsx.
(1) Use first row as column headers should be selected.
c. Import.
d. Choose your folder for the location.
e. Choose a Name; for example, “Global Bike Forecast Dataset”.
f. Description: “Global Bike Sales Data 2008-2020”.
g. OK.

Nitin Kalé & Nancy Jones © 2021 4 of 7


Practical Analytics Chapter 10 | Exercise 1 | Edition 3

h. Save. The Dataset will be saved with your other files and is available for use for other
analyses.

3. Create the Forecasting Model


1. Select Predictive Scenarios from the menu on the left side of the screen.
a. Select Time Series Forecast.
b. Choose your folder in which to save your forecast model.
c. Enter a Name such as “Sales Forecast”.
d. Enter a description: “Forecast of Global Bike sales”.
e. OK.
f. Settings:
(1) Enter a description.
(2) Under Time Series Data Source, select the dataset that you created in the previous
step.
(3) Choose Revenue for Signal.
(4) Choose Month_Year for Date.
(5) Number of Forecasts = 12 (12 months).
(6) Choose Currency for Entity. This will allow us to segregate U.S. and German sales.
(7) Leave the defaults for the Predictive Model Training. These settings are shown in
Figure 5.

Nitin Kalé & Nancy Jones © 2021 5 of 7


Practical Analytics Chapter 10 | Exercise 1 | Edition 3

Figure 5: Predictive Model Training Settings

Nitin Kalé & Nancy Jones © 2021 6 of 7


Practical Analytics Chapter 10 | Exercise 1 | Edition 3

g. Click on Train & Forecast. Be patient while the model is trained. Sometimes it takes a
minute or two. The results will be shown on three tabs or pages: Overview, Forecast, and
Explanation. Use these data to answer the assignment questions at the end of this
document. Be sure to expand the Explanation visuals to include all components.
h. Save the Sales Forecast Results in your folder. The Save Forecast function is circled in the
figure below. Provide a name of your choosing for the saved forecast. I named mine
“Sales Forecast Results”.

Figure 6: Save the Forecast Results

Question 1: Define MAPE. What is the significance of MAPE in this forecast?

Question 2: Take a screen shot of the visualization of your U.S. forecast.

Question 3: What are the forecasted sales for December 2021 in Germany? What are the forecasted
sales for December 2021 in the U.S.?

Question 4: Are there any signal outliers in the US data? If so, what are they? How might these
types of outliers affect this analysis?

Question 5: Use the model Explanation for the U.S. forecast and the information from the textbook
to explain what is triple exponential smoothing? How are alpha, beta and gamma
represented on the Explanation? Include a screen shot.

Nitin Kalé & Nancy Jones © 2021 7 of 7

You might also like