Pages

Monday, January 18, 2016

My Favorite: Linear Regression & Movies



Week 2 of the #MTBoS challenge: "My Favorite".

There are many lessons and activities that I have that could qualify as "My Favorite", but since I haven't blogged about this one yet, I'll go with it.

For the past couple of years, I used an activity I found on Yummy Math to help apply linear regression to a real world context.  I modify the activity just a bit by changing some of the movies to some of my favorites.  (If you haven't check out Yummy Math yet, I'd recommend it.  A lot of the site is free, but I'd recommend paying the $20 per year for full access.  The best thing about Yummy Math is that they're always updating their activities.)

Using Desmos, it's very quick and easy to create regression equations.  Students are asked to find some of the data from their own favorite movies, which completely draws them into the activity.

There is a lot of opportunity for rich discussion in the activity.  Some examples of really great discussions I've had w/ this activity:

  • Why are some movies way above or below the line of best fit?  What would cause that sort of behavior?
  • How does inflation / release dates play a role in this data?
  • Why do Disney animated movies tend to be below the line of best fit?
  • Why are sequels typically below the line of best fit?
  • Are most of the top-grossing films of all time (Titantic, Avatar, and now Star War 7) above the line of best fit?
  • When looking at a series of films (ex: Fast & the Furious), why do some fall above the line while others fall below?

Lastly, I love the activity because it leaves a bit of a cliffhanger.  Each year I do this, I wait and see what is the #1 movie in America during the past weekend and include that as part of my data set.  This past year, it was Hotel Transylvania 2.  At the end of the activity, I have students write down their prediction for the total gross amount for that movie.  Of course, there is no right answer at the time since the movie is still in theaters.  So we wait and occasionally check on it throughout the semester.  When the dust settles, we look back and see which student was closest to correct and award a prize.  



Title
Opening Gross
($ in millions)
Total Gross
($ in millions)
Opening Date
The Avengers
207.438708
623.279547
5/4/12
The Hunger Games: Catching Fire
158.074286
424.668047
11/22/13
Harry Potter & the Deathly Hallows Part 2
169.189427
381.011219
7/15/11
Jurassic World
208.806270
650.493056
6/12/15
The Sandlot
4.918712
32.434006
4/9/93
Toy Story 3
110.307189
415.004880
6/18/10
Dumb & Dumber
16.363442
127.175374
12/16/94
Frozen
67.391326
400.738009
11/22/13
Avengers 2: Age of Ultron
191.271109
458.924272
5/1/15
Major League
8.836265
49.797148
4/7/89








Hotel Transylvania 2
48.464322

9/25/15





2 comments:

  1. Thanks for posting this Mark. There is so much great data out there to hook kids on best fit functions. Your post makes me wonder if a data collection and this graphing activity might serve as bait/motivation for linear function interest initially. I usually do regressions as a re-visit to linear equations and an activity where we finally get to use what we learned about linear functions. Weaving in and out of regressions while learning the basics of linear functions might create the need for linear functions that has been missing in my class. I wonder if anyone else has interweaved the order with success or failure and written about it.

    ReplyDelete
  2. Sara,
    Thanks for the reply. I absolutely think you could use regression as a introduction to graphing, whether it be linear, quadratic, or any other type of function. I teach algebra 2, so I don't have the opportunity to introduce the concept of a linear graph to my students. It would be interesting to hear what algebra 1 / middle school math teachers would say.

    ReplyDelete