Description
You will study the Broadway Café case and perform Sales Analysis for it using Data Mining techniques.
Download and install the IBM SPSS Modeler (see separate instructions in D2L)
Read the Case: http://www.cohesioncase.com/CRM/AYK.html
(do not use the raw data .xls file though!)
Follow instructions below to run 3 types of Analysis and answer the following questions:
Association Analysis
What are two items with highest confidence that are closely associated each other?
What combination has highest support?
Classification Analysis
What types of sales is good for Music CD?
When is the best time to sell Gift Card?
Which item has highest refund?
Cluster Analysis
Is the Coffee good for promotional sale?
When the café plans to have promotional sales, what are items and when is the best time for the item?
Overall
What item has the biggest sales?
In order to improve sales, how can the café do?
Submit your work and research paper with screenshots to Dropbox
Participate in the Discussion
Submit analysis revision with concluding notes, with comments on what you learned and improved based on class discussion.
Data Mining with IBM SPSS Modeler
Download data file from D2L: CRM.Data.for.Class.xlsx
Run SPSS Modeler > Create New Stream (located top-right of the pop-up wizard) > OK
> Source tag (located at the bottom) > Click and drag Excel > Field Ops > Click and drag Type
> Right-click Excel > Connect > Click Type > double-click Excel > Data tag > File type: Excel 2007, 2010 (*.xlsx) > Import file: (locate the file) > Apply > OK
Right-click Type > Edit > Types > Read Values > Change Month to Categorical > Format > Double Click Month > Standard decimal places > Specify > change to 0 > OK > Apply > OK
Classification Analysis
Click Modeling > Classification on the very left at the bottom > Click and drag C5.0 > Right-click Type > Connect > Click C5.0 > double-click C5.0 > Fields > Select Use custom settings > (Choose one Dependent variable (Target) and at least one Independent variable (Inputs) to analyze. Here, trying to analyze Product based on Month and Sales Type.) > Target > click the icon on the right hand side > Click Product > Apply > Inputs > click the icon on the right hand side > select Month and Sales Type > Run > Double-click the model generated > Viewer > Try to interpret
Segmentation (Cluster Analysis)
Click Modeling > Segmentation on the very left at the bottom > Click and drag TwoStep > Right-click Type > Connect > Click TwoStep > double-click TwoStep > Fields > Select Use custom settings > (Choose variables to analyze. Here, trying to analyze Product, Month and Sales Type.) > click the icon on the right hand side > Click Product, Month, and Sales Type > Apply > Run > Double-click the model generated > View located bottom > change Model summary to Clusters > Click clusters > Try to interpret
Go back to main screen > Double-click TwoStep > Model > Specify number of clusters (change the number based on your intuition)
Association (Market Basket Analysis)
Click Modeling > Association on the very left at the bottom > Click and drag Apriori > Right-click Type > Connect > Click Apriori > double-click Apriori > Fields > Select Use custom settings > select Use Transactional Format > choose Customer ID and Product shown below
> Model > change it as the following image > Apply > Run