312541_en-US

liliconcom

Using Text Explorer to Extend Analysis

 

See how to:

• Understand terms used in text exploration- corpus, sentences, phrases, words, tokenizing, stemming, stop words, Document Term Matrix (DTM)
• Prepare data
  • Run Text Explorer and interpret basic word count report to find most common words
  • Use Stemming to group words (terms) with similar roots
  • Use Recode to consolidate synonymous terms
  • Replace related terms with a single phrase
  • Identify context specific stop words 
  • Customize Stop Lists to remove irrelevant terms from analysis temporarily or permanently
• Explore and analyze text  
  • Use Local Data Filter to examine responses for selected variables
  • Use Word Cloud to visually filter term occurrences
  • Review key terms in context
  • Create new variables associated with terms

• Briefly see how to perform Sentiment Analysis (JMP Pro)

• Use Text Explorer in a vacation case study example
  • Identify themes (JMP Pro)
  • Create groups (clusters) to see how often terms are used
  • Identify significant terms related to each cluster
  • Iteratively determine how many clusters are relevant to your problem
  • Save probability formulas to use as rules for other related data sets
  • Predict outcomes and use Document Term Matrix to convert significant text to numeric values to use in models
  • When geographic data are relevant and available, map text variables to related geographies
  • Use results to build a dashboard

Note: Q&A is interspersed beginning at ~time 29:30.

 

Resources:

• Mastering JMP video and files on Using Sentiment Analysis in JMP Pro
• Using the Regex Tokenizing Method

• Word Vector Embedding

• Text Explorer documentation and examples

Advanced Statistical Modeling