Cluster Analysis – Individual Rough Draft

Individual Rough Draft
Instructions
As an individual, you will create four cluster analyses with visualizations using two software packages: XLStat and Tableau. Be a data explorer. You will most likely need to try different methods to get the most interesting cluster.

Choose and prepare the data.
Open the dataset of interest.
Find two quantitative variables. You dont need to worry about them being correlated.
Copy your two variables into a new workbook.
Clean the data from any missing data points.
If your dataset is over 1000 observations, take a random sample of 300 to 500.
Save the new file.
Cluster Analysis 1. Use Tableau for k means Cluster Analysis.
Choose your own k you may need to do this a few times to decide the best clustering.
Create a main title and edit axis labels.
Record the sum of squares details (copy and paste to document)
Cluster Analysis 2. Use XLStat for k means Cluster Analysis.
Choose your own k from 2 to 10 you may need to do this a few times to decide the best clustering.
Choose Trace (W) or Determinant (W)
Create a scatterplot with the clusters in differing colors or characters.
Edit axes and axis labels.
Create a main title that includes the cluster method and linkage type.
Record the sum of squares details (copy and paste to document)
Cluster Analysis 3. Use XLStat for an Agglomerative Hierarchical Cluster Analysis.
Select Euclidean distance
Choose any linkage type.
Truncate the process between 4 to 10. Review the dendrogram (you may want to edit the number of clusters)
Create a scatterplot with the clusters in differing colors or characters.
Edit axes and axis labels.
Create a main title that includes the cluster method and linkage type.
Record the sum of squares details and dendrogram (copy and paste to document)
Cluster Analysis 4. Do your own thing!
Create a cluster with a different method, linkage and/or distance measure than before.
Create a scatterplot with the clusters in differing colors or characters.
Edit axes and axis labels.
Create a main title that includes the cluster method and linkage type.
Record the sum of squares details and dendrogram (if appropriate)
Paste each cluster analysis plot and the answers to the following questions in the individual discussion page.
Questions
Describe your two variables. What do they mean? Why did you choose them?
Which of the cluster analyses tells the strongest story?
In less than 300 words, tell the story of your clusters.
Attempt to name your clusters and describe them in context.
With what parts did you struggle/feel successful?

Last Completed Projects