Here is a simple tool we created from scratch, which points out the mean, median, and mode values from a randomly generated set of data:
Background:
The instructor noted that students are required to have taken a basic statistics course before starting his Sociology class, but they have often forgotten how to apply the concepts of mean, median and mode to a data set. He asked if we could create a tool that would show these values applied to a data set that the students might actually encounter during their sociology studies.
Details:
Note that the Median value(s) is highlighted in yellow on the graph. It is possible to have two values tied for median.
You can click the buttons next to “Surveys” to change out the data set. In theory the data displayed is close to what you will actually see when surveying the listed group.
You can click the “data” word atop the column of data, to sort the data from smallest to largest.
If you click the “randomize” button, it will generate a new set of data (but only for the currently selected survey). Unless you click randomize, each survey will retain it’s current data when you switch between them.
The chart generates a random number of entries (between 30 and 50). Then it ensures 20% are either 1 or 2; 60% are 3,4, or 5; and 20% are 6 or 7. This isn’t the best way to approximate a bell curve, but it was quick to implement.
Outcome:
All in all, this took about a week to create. We started talking about it on August 16th, and delivered this final version on September 9th, but there were lots of distractions (other projects) and vacation days mixed in. There was one graphic design overhaul in the middle, and some confusion about how to display the data. All in all, it seemed a quick and easy project.
The initial plan was to have each word literally point to the relevant block on the histogram. While we didn’t end up this explicit (due to communication hiccups between designer and programmer), we feel that the relationship still comes across pretty clearly..
There was a mild concern that randomly generated numbers won’t prepare students for real world results. We tweaked our random numbers so they they would adhere to more of a bell curve like shape, and decided this was close enough for the purposes of this simple exercise.
Initially there were two other surveys planned. But since we didn’t have an example data set to build from quickly, it was decided to just remove them from the exercise
We also agreed it would be cool to one day add a feature where you import an spreadsheet file (.CSV), and have the tool instantly give you the 3 resulting calculations. But we decided this wasn’t the point of this project, and have since moved on to other projects. (but maybe someday…)
Feedback:
If you have an idea for this project, or any questions, please feel free to post a comment.
