Friday, March 27, 2009

Quicksilver updates

This afternoon we had a very productive session with the BioSense BIC/Epi team to review the latest build of Quicksilver.

They gave a lot of good feedback that fell into two categories: easy changes (cosmetic) and more complicated. The complicated changes include modifying the map to color the cloropleth based on number of outliers per period (rather than the arbitrary count ranges currently used). We're going to work on this after Dr. Tokars' team is able to analyze the data to find the correct break points for what is significant and what is not.

The easy changes have been made into a tracker item. Here's the list as submitted by Peter Hicks and Steve Benoit:

  • Script error issues for every query requested

  • For clinical effects, listing in alphabetical order would be helpful

  • In legend scale, first option should be 0 (not less than 0), 1 to 10, 11 to 20, above 21. Currently, the categories overlap

  • Label what these numbers represent in the legend (visits, calls, etc. ?)

  • Adjustable legend option should be called “customize breakpoints or classification”. How to enter the customizable option is not clear or self-evident.

  • Queries are extremely slow

  • Calendar allows you to pick future dates – should have date limits

  • Colors in legend don’t align with colors on map.

  • A help section that describes the module and it’s components would be useful

  • Call out box when mousing over state shows a time series – the average line is not an average for the time period selected. The outlier box is not clear. How do you define an outlier?

  • Unable to understand this visualization and chart. We selected a region and got the call out box below. Was there 1 call or 797?

No comments: