• Stardust Quick Guide v1.05
  • 1. Introduction
    • Quick start
  • 2. Introduction to the PCA techniques
    • 2.1. Dimensionality reduction
      • 2.1.1. A first example: from 2D to 1D
      • 2.2.2. A second example: from 3D to 2D
      • 2.2.3. Viewing segments
      • 2.2.4. Going back to the Census-Income dataset
      • 2.2.5. Projecting the original axises inside the reduced space.
    • 2.2. Distance definition
      • 2.2.1. The K-Means Algorithm
      • 2.2.2. Automatic assignment of new customers to segments
      • 2.2.3. All the type of distances available inside StarDust
      • 2.2.4. A first example about the importance of variable “normalization”
      • 2.2.5. A second example about variables representing the same concept
    • 2.3. Summary
  • 3. TIMi Installation
  • 4. How to start « TIMi – StarDust module »?
    • 4.1. Introduction
    • 4.2. Using StarDust BEFORE a predictive analysis
      • 4.2.1. The DataSource Editor (Step 1)
      • 4.2.2. The Type Var file Editor (Step 2)
    • 4.3. Using StarDust AFTER a predictive analysis
  • 5. How To use « TIMi – StarDust module »?
    • 5.1. Viewing the population
    • 5.2. The central Toolbar.
      • 5.2.1.  Start new segmentation Analysis [FILE]
      • 5.2.2.  Open segmentation Analysis [FILE]
      • 5.2.3.  Save segmentation Analysis [FILE]
      • 5.2.4.  Parameter Windows
      • 5.2.5.  Camera move and rotation Mode [MOUSE MODE]
      • 5.2.6.  Mouse Pick Mode [MOUSE MODE]
      • 5.2.7.  Invert selection
      • 5.2.8.  Auto-rotate 3D view
      • 5.2.9.  View one Var Distribution [CHART]
      • 5.2.10.  Compare variable distribution for many variables [CHART]
      • 5.2.11.  View Summary Table [CHART]
      • 5.2.12.  View final report
      • 5.2.13.  Keep Display Aspect Ratio?
      • 5.2.14.  Show the Segments
      • 5.2.15.  Compute new segmentation and shows it!
    • 5.3. Setup the 3D display
      • 5.3.1. The (Meta-)Group tab
        • 5.3.1.1. Creating a new group based on the current selection
        • 5.3.1.2. Configuring the colour and motif of the 3D points
        • 5.3.1.3. Manipulating meta-Groups
        • 5.3.1.4. Creating Groups based on modalities of variables
        • 5.3.1.5. Creating Groups based on the segments
      • 5.3.2. The graphic tab
        • 5.3.2.1. (Meta-)Group selection for Display
        • 5.3.2.2. Axises
        • 5.3.2.3. Item Labels
        • 5.3.2.4. Display the original axises inside the 3D view.
    • 5.4. Configuration of the PCA analysis
      • 5.4.1. Detecting and removing outliers using the PCA
      • 5.4.2. “Zooming” on the data using the PCA
    • 5.5. How to create a segmentation?
      • 5.5.1. Definition of the Distance to use inside the segmentation
      • 5.5.2. The parameters of the K-Means algorithm
        • 5.5.2.1. Control 1: Run K-means
        • 5.5.2.2. Control 2: Sampling
        • 5.5.2.3. Control 3: Energy
        • 5.5.2.4. Control 4: Random seed
        • 5.5.2.5. Control 5: Retries for seeding
      • 5.5.3. The Ward’s algorithm
        • 5.5.3.1. How to find the right number of segments?
        • 5.5.3.2 The different variations around the original Ward’s algorithm
      • 5.5.4. Setup the display of the segments
      • 5.5.5. Saving the segments in a CSV file.
        • 5.5.5.1. The “Save Core” option is enabled
        • 5.5.5.2. “Save the Distance” to the nearest centroid.
      • 5.5.6. Exporting and using segmentation models.
      • 5.5.7. Explaining the segments from a business-point-of-view
  • 6. Virtual Reality (VR) display
  • 7. Conclusion