Reference no: EM133096429
Time Series Analysis
Objective: The objective of this exercise is to use time series analysis to forecast GBI sales
Activities
• Import and prepare data
• Apply data mining algorithms
• Configure forecasting models
• Create data visualizations
• Analyze and interpret output from models
• Publish results
SCENARIO
GBI sales revenue declined in the US during the 2008 financial crisis and although it has picked up a bit since, Nina is interested in forecasting sales revenue for the immediate future, at least one year from the date for which data are available. She will use time series analysis for forecasting.
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. Launch SAP Predictive Analytics
2. Click on Expert Analytics → Expert Analytics
3. Acquire the data
a. Click on File → New
b. Choose Microsoft Excel → Next
c. Choose the File sales_transactions_E12_1.xlsx (Hands-on- 4_Sales_Transactions_Data.xlsx)
d. Open.
e. See the preview of the data
f. Click Create
g. After data are acquired, Click on Visualize tab
h. Plot a line chart of revenue vs Year and Month as shown in Figure 1
4. Save the PA file
5. You will now use triple exponential smoothing to forecast the sales revenue for year 2017 based on monthly sales figures from 2007 through 2016.
a. Click on Predict tab
b. Under Data Preparation, double click Filter.
c. Configure Settings on the Filter icon.
d. Create a row filter that filters revenue for US only by filtering currency to USD as shown in Figure 2. [Currency]=='USD'.
e. Under Algorithms, double click on Triple Exponential Smoothing.
f. This will connect the Triple Exponential Smoothing algorithm to the filtered data source.
g. Now Configure Settings for the Triple Exponential Smoothing icon.
h. Configure the settings as shown in Figure 3.
i. Click Done.
6. Running the algorithm
a. Click Run.
b. After the model runs, Click OK.
c. You will see the data grid view of the forecasted values of sales revenue.
d. Click on Trend Chart.
e. You see a combined column and line chart showing the historical sales revenue, fitted curve of revenue, forecast for revenue for 12 months in the future (Figure 5).
f. Save your PA file.
1. Paste a screen shot of the results of your forecast into a Word document. Label it with the alpha, beta and gamma settings.
2. Explore the results tab to find the R-square factor. What is the R-square factor telling you about the model? Also look for Goodness of Fit. Comment on it as well.
HANDS-ON WORKSHOP
Time Series Analysis
NITIN KALÉ, UNIVERSITY OF SOUTHERN CALIFORNIA NANCY JONES, SAN DIEGO STATE UNIVERSITY
EXERCISE 1 - Objective
The objective of this exercise is to use time series analysis to forecast GBI sales
Activities
• Import and prepare data
• Apply data mining algorithms
• Configure forecasting models
• Create data visualizations
• Analyze and interpret output from models
• Publish results
SCENARIO
GBI sales revenue declined in the US during the 2008 financial crisis and although it has picked up a bit since, Nina is interested in forecasting sales revenue for the immediate future, at least one year from the date for which data are available. She will use time series analysis for forecasting.
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. Launch SAP Predictive Analytics
2. Click on Expert Analytics → Expert Analytics
3. Acquire the data
a. Click on File → New
b. Choose Microsoft Excel → Next
c. Choose the File sales_transactions_E12_1.xlsx (Hands-on- 4_Sales_Transactions_Data.xlsx)
d. Open.
e. See the preview of the data
f. Click Create
g. After data are acquired, Click on Visualize tab
h. Plot a line chart of revenue vs Year and Month as shown in Figure 1
4. Save the PA file
5. You will now use triple exponential smoothing to forecast the sales revenue for year 2017 based on monthly sales figures from 2007 through 2016.
a. Click on Predict tab
b. Under Data Preparation, double click Filter.
c. Configure Settings on the Filter icon.
d. Create a row filter that filters revenue for US only by filtering currency to USD as shown in Figure 2. [Currency]=='USD'.
e. Under Algorithms, double click on Triple Exponential Smoothing.
f. This will connect the Triple Exponential Smoothing algorithm to the filtered data source.
g. Now Configure Settings for the Triple Exponential Smoothing icon.
h. Configure the settings as shown in Figure 3.
i. Click Done.
6. Running the algorithm
a. Click Run.
b. After the model runs, Click OK.
c. You will see the data grid view of the forecasted values of sales revenue.
d. Click on Trend Chart.
e. You see a combined column and line chart showing the historical sales revenue, fitted curve of revenue, forecast for revenue for 12 months in the future (Figure 5).
f. Save your PA file.
g. The forecast depends on the choice of alpha, gamma, beta parameters for triple exponential smoothing. You can Click on the Designer tab and then reconfigure the setting for the Triple Exponential Smoothing algorithm. Choose different values for alpha, beta, and gamma in the Advanced tab. Then rerun the model.
3. Paste a screen shot of the results of your forecast with the new alpha, beta and gamma into your Word document. Label it with the alpha, beta and gamma settings and explain how the forecast has changed.
Attachment:- Time Series Analysis.rar