Aug 13, 2011

Posted by in Visualizations | 37 Comments

Visualizing classroom seating habits

Edit (August 29, 2011): the visualization has been updated, thanks for all your comments.  You can find out more about it on Behance.

I’ve been working on the following interactive visualization for the past couple of weeks, primarily since I thought it would make for an interasting dataviz and also as an excuse to learn D3.  It captures the seating habits of a class of graduate students attending the class 15.514 in the summer of 2011.  This is still a work-in-progress.  I’d love to hear your comments.

http://skyrill.com/seatinghabits/

There will be a tutorial that will take you through how it was developed and the toolkit used for it soon, so be sure to look out for it.

Have a great weekend!

 

  1. really nice idea, great one!

  2. Looks great…will keep checking back for progress

  3. I can see that one person reeeeally likes their seat :-)

  4. Neat (and pretty) visualization, however without having swatches of the color associated with each student next to their name (i.e., making it a legend), or providing names on mouseover, the only way to know who is who is to select them one at a time, making it difficult to know what any visually identifiable trends mean.

    • Thanks, that’s a great suggestion. How does it look now?

      I think it would also be useful to highlight names when the mouse hovers over their respective circles since in some cases, the colors look very similar. They’re currently auto-generated by d3 (you might have noticed that after 20 colors, they repeat).

  5. I’d be interested to see how this corresponds to grades.

  6. Very cool.

    A time based visualization could be interesting if you have that data, to see if people migrated around the classroom together.

    Also, using wrapping the names in: *label for=”check_box_id”* */label* where you replace * with open close HTML tags, would allow the names to be clickable.

    • Thanks for the suggestion. Done and updated!

      The other suggestion would be really interesting; that’ll be worked on next: the idea, as you point out, is to be able to select a subset of students and then see how they move around the classroom during the semester.

  7. I’d like to see where the door is in the room in relation to the seats… :)

    • There are two doors on the left: one at the very front and one at the very back.

      I think adding some abstractions for doors, the blackboard, etc. would be useful.

  8. Fantastic. Would also be awesome seeing which students like sitting next to each other a lot, as well — by proximity of each seating!

    • Yep, will hopefully have that done during the coming week (or even the weekend if the hurricane actually hits Boston and I’m stuck indoors :) )

  9. I wonder to what extent seat preference depends on the time a student arrives to class. I’d imagine those students that are habitually late have little choice as to where to sit, whereas those who always arrive early can have their first pick of seats.

    • Or whether or not some of the students attended the class right before it, which is the case on some days of the week.

  10. It looks like the people who prefer a single seat are monopolizing one area; one would like to know why — so placement of doors, blackboard, teacher etc.

    Ditto, there seems to be some type of exchange or competition between just a few people for the front row on the left.

    Definitely seeing the time sequence for when each seat is filled (by person) would be interesting, as per comment by Talie and others.

    Anyway, it’s a nice visualization, easy to navigate and interesting to think about — especially if you add the grades as per another comment.

    • Those are some interesting observations.

      Check the updated visualization uploaded a few minutes ago. There’s a new button now that will allow you to go through the time series data for an arbitrary set of students.

      Thanks for the feedback.

  11. When I was in grad school a fellow grad student collected data on where faculty, grad students, and undergrads sat during weekly colloquium lectures. She then normalized their position relative to one particular faculty member who seemed to think himself the center of the universe, and gave a very entertaining presentation on it at an annual grad student awards ceremony. Not quite as informative as your visualization though.

    It would be fun to bring in data on which students asked the most questions. I wonder if the same student asks more questions when they are in some positions than when they are in others? Maybe the same student asks more questions when they’re in the front row than when they’re in the back for example. If so, is this causal, and which way does the causality point? Perhaps sitting at the front of the class is an “independent” event and leads to students feeling pressured to ask more questions. Alternately, maybe when a given student is in a good mood for asking questions, they sit at the front, and thus any correlation is confounded by an unseen mood variable.

    Also it seems natural to wonder about how the relationships between students might influence their seating choices. Some sort of clustering analysis on the data might reveal that certain students sit together, even though they move to different areas of the room. Perhaps you could draw lines between students, with that line faded to represent how strongly the two students are correlated. To prevent the graph from being cluttered, you could set some threshold below which nothing was plotted. This would be particularly entertaining during the time-series animation.

    • Thanks, Hig, for the feedback and wonderful insights! Those are all intriguing points and would certainly enrich the visualization.

      I like the clustering analysis idea in particular and will see if I can make the time to work on it next week.

    • when i watched this the first time my eyse picked out a pattern with Wayne and Sofia that’s fun to watch. then you can look at Mary and Lila. I could only guess at the gender, but having been a grad student before, I can’t think it is irrelevant. Another item of interest is to see the order that students sit down and time of entry. Does Mary take Sofia’s place next to Wayne one day because Sofia’s late. Well, I’m sure it’s more exciting to think about the possibilities than what was actually happening….

      Overall this is excellent and a key way to see that is how many people want more and more info.

  12. Great idea – it would help to identify the front of the classroom where the instructor is!

  13. Dig it. I need to know where the teacher stands. Also some axis to mark front/back embedded would make the figure more self-sufficient (and therefore more awesome-er).

  14. Saw this on Flowing Data. What person or object is directly in front of the third seat on the right? Seems that no one sits in that seat ever so I figured something is either blocking the view or students feel uncomfortable being that close to the teacher.

  15. Ive always wondered how long it takes a class to “settle down” in its seating. The second day’s seating is seldom like the first, as you spot friends, find out your neighbor smells, etc. The third day is slightly more stable. And so on, until everyone has their same seat that they go to. I wonder how long it takes? Does it depend on the class subject? Size of the room? Age?

  16. Love to know some of the details:
    male vs female
    major (studies)
    age, approximate

  17. Excellent visualization! I’ve always joked that physics students tend to behave like fermions… spacing out and never occupying the same area until the others are filled. Beyond the visualization, it would be interesting to see what sorts of models are in fact appropriate… and as suggested above: how do the various “boundary conditions” play into it.

    Thanks for sharing this!

  18. Assuming the front of the classroom is at the bottom of the infographic…

    It looks like no one favors the seats furthest to the right. Those seats are often coveted by left-handed people like myself. Is everyone in the class right-handed?

  19. Hi, this is fantastic guys. Is this a classroom with individual tables? I want to do something like this but in a lecture theatre to see simple patterns in how people habitually tend to choose the same thing, almost without thinking (i.e. autopilot). I wonder how often this type of simple experiment has been documented before? Thanks for sharing I will be keen to learn how to simulate this kind of representation :)

  20. Greetings from Brisbane, Australia. My name is Nicky. I am a nineteen year old business student. I found this blog when I was doing some research for my school assignment. Thank you for your help. I now have a lot more direction. :)

  21. The website works just fine with Internet explorer 9 & 10. Maybe time to shake off that warning page.

Leave a Reply