Homework #2:  JMP Data Analysis Tools

The goal of this homework is to experience and practice the use of visualization tools, accustom yourself to seeing data in new ways, discover something new in a data set, and think critically about visualizations.

Your assignment is to find some interesting data, explore it using some visualization tools, and write a short report.

Data:  Find some data of interest to you.  The data should be large and complex enough that visualization is needed.  It should have at least 500 items and at least 5 attributes.  Here are some data ideas:

• network intrusion detection (data)
• scientific experiments (data, description)
• bioinformatics (data, description)
• census data  (census.gov)
• stock market data
• sports data

Visualization:  Examine your data using the SAS JMP tool available in McBryde 104c (also available as downloadable demo at that link). It can accept data in a variety of formats including Excel spreadsheets or cut&paste. It has extensive online documentation. It includes several of the multi-dimensional data visualization techniques discussed in class, all linked together. In your analysis, you should make use of at least 3 of the following techniques:

• Histograms (JMP Starter | Graph | Histograms)
• Parallel coordinates (JMP Starter | Graph | Parallel Plot)
• Scatterplot matrix (JMP Starter | Graph | Scatterplot Matrix)
• Heat map (JMP Starter | Graph | Cell Plot)
• K-means cluster (JMP Starter | Graph | K Means Cluster) -- this is a special case in that it does not have a visualization of its own, but rather creates a cluster-based coloring that is shown in the other visualizations.
• another tool of your choice.

Report:  Write a report about your data findings and about the visualization tools. Specifically answer these 3 questions:

1. Data:  topic, number of items and attributes, and meaning of the items and attributes. Brief.
2. Data findings:  List 3 interesting, non-trivial findings in your data.  e.g. Did you find answers to questions you had about the data? What relationships exist in the data? Anything surprising? Back up your claims with pictures (Alt-PrintScrn captures a screenshot that can be pasted into word processors).
3. Compare and critique:  Compare the techniques you used within JMP in terms of the visual mapping, scalability, and insight provided (user tasks supported). This section is most important and should take a full page.

Format: 2 pages, 11pt font, single spaced. Attach pictures as additional pages.

What to hand in:

• Hardcopy due in class, Mon Sept 15.

Access:

• Mcbryde 104c:  Beth will post hours of availability.  It may also be open during the day, but is locked and inaccessible otherwise.  So do NOT wait until the last minute.
• Please be careful with equipment in 104c, and leave everything as you found it.