Explanation.Txt History

Hide minor edits - Show changes to output

November 27, 2006, at 02:15 PM by -
Added lines 1-33:
Hi, Saul. Attached below is the figure I used in my CSCW Annotations talk
that you said you'd like to get a copy of.

I've also attached another plot using the same method, except the plot
focuses on annotators instead of documents. Here's the explanation for both

pianoRoll_documents: Each row is a document, and each dot represents a day
on which at least one annotation was made to that document. Documents are
sorted in order of date on which the document was first annotated. A few
interesting things can be derived from this plot. First, you can see the
rate at which new documents are being annotated by looking at the slope of
the left-most points. Interestingly, this slope also matches up fairly
nicely to significant deadlines in this product group's schedule. Second,
you can see gaps in the plot for holidays (July 4th, late November, late
December). Third, as mentioned in my presentation, documents are annotated
over a very long period of time, sometimes as long as several months.

pianoRoll_annotators: same idea as the previous plot, except this time each
row represents a single person. The plot is sorted such that the people who
made the most annotations are at the top and the people who made the fewest
annotations are at the bottom. We constructed this plot to see if the
different classes of annotators used the system at different points in time.
In particular, we were trying to see if the one-time users all tried the
system early in the process when it was first announced and then never used
it again. From this plot, it doesn't look that way.

Both plots were done in excel. Marc Smith should get the credit for the
idea of displaying the data in this way.

Let me know if you have any other questions.