Skip to content

Clustering Behavioral Profiles with Cluster 3.0

Behavioral profiles generated using Excel Template B are ideal for clustering, which can identify patterns in your data that were not easily discernable. Cluster 3.0 has a variety of clustering methods and customazible paramaters. In this tutorial we will specifically be using hierarchical clustering with euclidean distance and complete linkage.

Materials

Cluster 3.0

Java TreeView

Installing Cluster 3.0 and Java Treeview

  1. Download the version of Cluster 3.0 that matches your computer (e.g. if you are using a mac, download the macOS version) by clicking on the link from the list of materials.

  2. Open the .exe file and install the software.

  3. Download Java Treeview by clicking on the link from the list of materials.

  4. Extract the downloaded file (right click on the file -> extract here) and click on the TreeView launcher to open Java TreeView (1).

    1. If you can't open the application, install Java on your computer.
    2. Click on the A to select every value on column A.

Preparing your Data

  1. Open your data, analyzed in Template B.

  2. Go to the Differences tab and copy all the groups.

    Copy the groups

    Experimental Groups

  3. Go to the Text for cluster analysis tab and paste the groups in the first column by using the paste values option (i.e. right click -> paste values).

    Paste the groups

    Experimental Groups

  4. Delete any extra groups.

    Clearing extra data

    Deleting Groups

  5. Delete the rest of the extra data by selecting column A (2) , then clicking on Find and Select -> Go To Special... -> Select Blanks and click Ok.

    1. If you can't open the application, install Java on your computer.
    2. Click on the A to select every value on column A.
    Select extra rows

    Selecting Rows

  6. Delete the extra rows by clickin on Delete -> Delete Sheet Rows.

    Delete Rows

    Deleting Rows

  7. Save the Excel sheet as a tab delimited file by going to File -> Save as -> Text (Tab Delimited) (.txt). Click Ok when the pop-up asks if you want to save a single sheet.

    Save settings

    Save Settings

Hierarchical Clustering

  1. Open Cluster 3.0.

  2. Open the .txt file you saved in the previous step.

    Open file

    Open File

  3. If the file loaded correctly, it will look like this:

    Loaded file

    File

  4. Go to the Hierarchical tab and enter the following settings, then click on Complete Linkage.

    Clustering Settings

    Settings

  5. Once the program is done, it will say Done Clustering at the bottom. It will generate two files with .cdt and .gtr extensions. These are located in the same folder as your .txt file.

Visualizing the Clusters

  1. Open Java Treeview

  2. Open the .cdt file you generated in the previous step.

    Loading CDT file

    CDT File

  3. You now have a hierarchical cluster of your data! Feel free to play around with the color settings.

    Clustering Results!

    Cluster Results

    Suggested Settings

    Pixel Settings: Zoom = 40 x 20, Contrast = 25

    Font Settings: 20