Difference between revisions of "Rdviz-uncertainty"

From Responsible Data Wiki
Jump to: navigation, search
(Questions to ask oneself about visualizing UNCERTAINTY)
(Contributors)
Line 26: Line 26:
  
 
== Contributors ==
 
== Contributors ==
 +
@vm_wylbur
 +
@sxedwards
 +
@mikel
 +
@shugg1elee
  
 
== Resources (we <3 links!) ==
 
== Resources (we <3 links!) ==

Revision as of 01:21, 23 January 2016


Questions to ask oneself about Visualizing Uncertainty

- What does your viz purport to represent in the real world?
- Is your visualization introducing uncertainty to forgo conflict? In other words, are you being "more vague" in order to be "less politically-charged"?
- Does your error overwhelm your argument? Is your argument robust to error? Ideally your visualization should show margins of uncertainty or error in addition to your findings.
- Who is your audience? Your relationship to your audience determines your trade-off between certainty and precision.
- The knowledge of the uncertainty of your data is just as important as the data... in fact it is the data.
- What are you assuming when you compare unrepresentative datasets visually

Solutions for Visualizing Uncertainty

Below are some ways to visualize uncertainty in your data / research - Identify alternate conclusions
- Aggregate to increase certainty (at the cost of precision)
- When your data creates much uncertainty, create relative visualizations or use grading
- Accept the uncertainty, highlight the uncertainty, celebrate the uncertainty

The Assumption

For the purposes of this afternoon's exercise we operated under the assumption that we can measure or assess uncertainty. this of course is our way of also acknowledging that there is uncertainty that is very hard, or even impossible, to measure.

Audience

Personas, use cases, context

Next steps

Contributors

@vm_wylbur @sxedwards @mikel @shugg1elee

Resources (we <3 links!)