Friday, November 26, 2010

Toasting Time is Not Linear

Dan Meyer just posted some data on his dy/dan blog about toaster settings vs. toasting time.  You can see the video posted below.

Meyer — Toaster Regression from Dan Meyer on Vimeo.

I just ran the toasting times through Microsoft Excel 2010.  The raw data is below.
Raw Data

I graphed the data and used trendlines to analyze the data.  The first graph shows the linear regression.
Linear Regression


The next graph is the exponential regression.
Exponential Regression


Finally, here is the quadratic regression.
Quadratic Regression
Based on the correlation values, the best model is the quadratic.  The exponential is second best.  The linear is the third best.  However, all are good representations.

Also on Dan's blog post, the darkness of the toast is estimated.  The graph, with exponential regression, is below.