This week in my social media innovation class we covered pagerank as the foundation for search engine optimization. In preparing for class, I couldn’t locate any activities that illustrate the key concepts of page rank… so I created one myself!
Collaborative Filtering Voting Activity (a la pagerank)
The process is a little bit complicated, but it worked out well. I used Lego-like blocks as a tangible symbol of “voting,” akin to the votes a website receives through in-bound links.
The high-level process went like this:
- I quickly organized students the 55-60 students into 16 (3-4 person) teams, assigning each team a number.
- The students selected a team Reader, Recorder and Voter(s). I gave the Voters 10 blocks per team.
- The students completed short simple task with an outcome that could be readily assessed by other teams (in this case, asking students to develop a 5-7 word tagline for our class).
- They then went through two rounds of voting for the best tagline. The Voters roamed the room talking with Readers. Voters handed out 1 or more blocks to the team of their choice. The Recorder noted the team number the blocks were received from.
- Round 1
- In the first round every block counted equally: as one vote.
- At the end of the round each Recorder announced their vote total. I entered these into a spreadsheet that the class could see. I rank ordered the teams by total votes and then assigned each team a weight of 1 to 5 based on that rank (with a 5 point weighting for the most popular and 1 point for the least).
- The Recorders then turned over their team’s blocks to the Voter for Round 2. (One team received zero votes in Round 1: I gave them a block to use for Round 2.)
- Round 2
- In the second round the value of a block varied depending on the weight for the team that gave it.
- Voters, Readers, and Recorders repeated the voting process.
- At the end of the round, Recorders calculated the weighted total for their teams votes.
- We did another roll call, identified the winning team and did a debrief.
The entire activity took about 40 minutes. The winner of Round 1 dropped near the bottom in Round 2. The winner from Round 2 was a team that finished in the middle after Round 1.
This exercise is not intended to teach students the Pagerank algorithm! Instead, it helps students understand the managerial implications of the algorithm for marketing a website.
- Students learn that in-bound links, not out-bound ones, determine pagerank.
- Students learn that not all in-bound links are the same–some count more than others.
- Students observed there is more to a high rank than good content (in this case, a course tagline). Additional factors play into relatively popularity. Most noticeably in this case, the group physically situated in the middle of the back of the room was hard to get to and, unsurprisingly, finished with the lowest vote totals.
- For a smaller class, I might try a third round of voting or repeat the exercise with a small variation (either a new task or some voting limitations).
- If I had fewer teams I would also provide fewer blocks to vote with.
- Instead of re-using the blocks from Round 1 as the number of Round 2 votes, a more realistic scenario is to hand out a second set of blocks so everyone has equal votes.
- Also, rather than determining the winner solely based on the Round 2 totals, I would probably add together scores from Round 1 and Round 2.
I am a big believer in the use of in-class activities to promote learning. Activating multiple pathways–both physical and intellectual–reinforces both learning and recall of key concepts.
Do you have any favorite in-class exercises for under-graduate or master’s students?
Feel free to adapt my handout for this exercise (Prof Johnson Pagerank Activity). If you do, please let me know how it goes for you!