I used the project gallery for my dataset, recording project names, views, word lengths, number of people involved, and word/character count from all projects in this, the co'llusion, and the Ignis fatuus project pools. I thought I'd then try some more subtle Fox News techniques.
The dataset can be viewed here: http://bit.ly/1Wph9mrMy goal for this project was to skew the data as much as possible to show that traditionally "better" IDeATe Gallery projects, specifically in this class, do not fare as well. I chose this because the opposite is precisely true; those projects which are new, interesting, and informative bubble to the top of the view/applause count and fare much better, from a popularity standpoint, than others.
Obviously, Applause means much more than Views when it comes to projects, for many reasons: users can view their own work and simply refresh the page to rack up views, while only other users can applaud at most once in a project. So I decided to completely ignore applauses, attempting to show that projects with fewer and smaller words typically did better.
I decided not to lie with my initial data, but only with its visualization. Aggregating information in interesting ways led to very misleading charts. For one, I used select subsets of the (already small) sampling of projects to skew my results in pointed ways. I also flipped axes to make things seem more prominent than they were, cut off graphs to show only what I wanted users to see, and generally sum the information in meaningless ways.
To view the charts, click "See Data" floating in the top left. Each chart involves a particular lie, and to see how each chart is generated, click the link in the "Dataset" section.
To show that I know what I'm talking about, I used a little CSS to clean up this page. I've kept the write-up incredibly short up until this point, and I only let the user see my charts after I get rid of all the other clutter on the page. This has an apparent effect: as of 1:00am on the day of the deadline, I have over 3,000 views and 90% of the class-wide applauses!
Well, actually, that was the final lie. After posting my initial results, I added a little script in the documentation that fakes views and applauses, causing the numbers themselves to be skewed. Now, even with a "fair" representation of the data, I can get my point across!
Overall, I think that I did a pretty good job of skewing the numbers. Given some more time, I'd love to construct the graphs from scratch (using HTML/CSS on this page) rather than just using the Google Sheets images. I learned quite a bit about the front-end of the IDeATe gallery as well as that views are actually much more evenly distributed than I had thought.
You can upload files of up to 20MB using this form.
