Mike Freeman, from the Institute for Health Metrics and Evaluation at the University of Washington, submitted this interactive visualization for feedback and advice. He is requesting comments on three specific areas (but, sorry, the data are not available):
1. Scales: When you switch categories (sex, for example), the scale changes to accommodate the new range. We put a lock icon to the left of the y-axis, but this only allows locking across years. Is there a good way to design scale schemas such that people don’t misinterpret updates when both data and scales change? Perhaps transition the data on the current scale, then change the scale.
2. Uncertainty: There is notable uncertainty around some of the estimates—in fact, I worked on another visual designed specifically to represent the uncertainty in the estimates. Any thoughts on a good way to represent uncertainty in a stacked bar chart?
3. Simplicity versus flexibility: We wanted users to be able to view the data however they wanted—to choose the metric, unit, sex, location, and year. Unfortunately, I think the number of controls can overwhelm users and there is nothing that simply states what the user is actually viewing (space was limited). What approaches have people used to offer simplicity to a broad audience while providing flexibility to advanced/interested users?