It can be difficult to extract patterns with complex sets of data. Visualization tools can be used to identify relationships that would not be otherwise evident.
Students seem to always sit in the same seats in the class. Data visualization provided an opportunity to examine the seating habits of students in the author’s financial accounting class throughout the semester. This involved:
• gathering data
• transforming the data into the right types of data sets
• loading the data into the visualization tool
• running the tool and observing the results
Where every student sat for each lecture was recorded in a spreadsheet. The data were transformed into two representations:
• a static set that listed each student, the seats he or she sat in and how frequently the student sat in each seat
• a dynamic set that showed a time-series for where students sat throughout the semester
The data sets were loaded into Data-Driven Documents (D3), a powerful visualization toolkit developed by Stanford’s Visualization Group. The visualization renders a circle for each student-seat pair. Hence, a student who sat in a total of four seats during the semester would have four circles rendered on screen, each at the coordinate corresponding to the seat’s location. Each circle’s diameter is determined by the number of times the student sat in the seat — bigger circles for more frequent use.
Figure 1. The visualization can show a time-series animation
of a particular student's seating behavior. This shows a static
view of one student's behavior throughout the semester.
Results and observations
The visualization revealed that there are three groups of students, in the following order from most to least:
• those who preferred to sit in the same zone, or set of seats
• those who moved seats, showing no preference
• those who chose to sit in the same seat
For the SDM ’11 cohort in this one class, most students didn’t sit in the same seats but rather in the same zones.
Suggestions for extending this project include having a clustering algorithm to investigate correlations between where people sit and the following:
• relationships between students
• what time they arrive to class
• personality type
• whether or not they have a class immediately before
• final grade
Ultimately, this experiment showed that transforming data via analysis and visualization is a great way of seeing patterns that would otherwise not be apparent.
About the Author
In June, Ali Almossawi will join Mozilla Corporation (Firefox) as a metrics engineer. He is also cofounder of the design firm Skyrill.com. Almossawi has a master’s degree in software engineering from Carnegie Mellon University. He recently completed the SDM master’s program.