Using JMP® Pro for Enhanced Marketing Analytics and Consumer Research

Presenter: Jian Cao

Process and Visualize Free-Form Text, Prepare it for Analysis and Interpret Results

See how to:

  • Launch Text Explorer (5:25)
    • Identify Terms as the smallest pieces of text, similar to words in a sentence; Phrases as short collections of terms; and Documents as collections of words in a table cell
    • Use Stemming to combine words that have common roots
    • Use Regular Expressions to find patterns in words
    • Evaluate and visualize prevalence or Terms and Tokens (number of words)
  • Interactively examine and use Word Cloud to visualize text and interactively see full text (Documents) related to specific Terms (9:00)
  • Add new variables that summarize responses, for example, opinions collected in a survey (11:00)
  • Use JMP Pro Latent Class Analysis to group top Terms into clusters and Discriminant Analysis to show the relationship of Terms to a response (12:49)
  • Understand how JMP uses Latent Semantic Analysis and a partial Singular Value Decomposition (SVD) to reduce Terms into a manageable number of dimensions for analysis (13:48)
  • Examine the Document Term Matrix (DTM) 
    • See how DTM changes Terms to numeric values that can be used in analysis
    • Clean up DTM to remove empty cells
  • Cluster and weight similar Terms and Phrases (18:48)
  • Group related Terms into latent groups (Topics) and examine the Terms most and least likely to be associated with each Topic
    • Use Topic information to build predictive models (26:25)
  • Understand and run Association Analysis (Affinity Analysis, Market Basket Analysis, Pattern Recognition) (33:00)
    • Understand and specify Support, Confidence and Lift values to identify Association Rules (relationships of Terms)
    • Examine and interpret Confidence and Consequence for each identified association 

Note: Q&A included at times 29:36, 31:07 and 31:23.

Resources

Back to Top