VISUALIZING UNCERTAINTY Fall 2017 Mac Hill VISUALIZING UNCERTAINTY - PowerPoint PPT Presentation

Thesis Proposal VISUALIZING UNCERTAINTY Fall 2017 Mac Hill

VISUALIZING UNCERTAINTY 2 DEVELOPING A VISUAL VOCABULARY FOR UNCERTAINTY How can the design of information visualizations commonly found 
 in news media coverage incorporate representations of uncertainty in situations that depict predictive data to facilitate non-expert decision making about current events?

VISUALIZING UNCERTAINTY 3 DEFINING UNCERTAINTY Incomplete or imperfect knowledge arising from a variety of factors including: measurement precision, completeness, inferences, disagreement, and credibility. Skeels, Meredith, et al. “Revealing Uncertainty for Information Visualization.”

VISUALIZING UNCERTAINTY 4 TYPES OF UNCERTAINTY Disagreement Completeness Inference Precision Uncertainty can be introduced during the collection, analysis, 
 or presentation of information. Multiple types of uncertainty can 
 be present in a single information visualization. Skeels, Meredith, et al. “Revealing Uncertainty for Information Visualization.”

VISUALIZING UNCERTAINTY 5 “Uncertainty is a fact of information; 
 all information contains uncertainty” A. M. MacEachren, et al. "Visual Semiotics & Uncertainty Visualization: An Empirical Study."

VISUALIZING UNCERTAINTY 6 WHY INCLUDE UNCERTAINTY? The presentation of data has an impact on its meaning and its usefulness for decision making. Leaving out uncertainty provides a skewed and incomplete picture of the information being visualized. Edward Tufte. Envisioning Information

VISUALIZING UNCERTAINTY 7 MISSING UNCERTAINTY Visualizations like this one leave out the uncertainty inherent in predictive and polling data. In this case, the margin of error is larger than the lead one candidate has over another.

VISUALIZING UNCERTAINTY 8 Information visualizations in mass media often leave out representations of uncertainty. Instead they rely on percentages to convey doubt. UNCERTAINTY IN MASS MEDIA FiveThirtyEight.com, “Who Will Win the Presidency?”

VISUALIZING UNCERTAINTY 9 The raw data from the previous visualization looked more like this, with 1,106 polls combined together and several showing Trump winning. FiveThirtyEight.com, “Who Will Win the Presidency?”

VISUALIZING UNCERTAINTY 10 “there are real shortcomings in how American politics are covered, including... 
 a failure to appreciate uncertainty” Silver, Nate. “The Real Story of 2016”. FiveThirtyEight.com

VISUALIZING UNCERTAINTY 11 STATISTICAL METHODS FOR VISUALIZING UNCERTAINTY Error Bars Box Plots Violin Plot Confidence Intervals Blur Statistics has some methods for visualizing uncertainty. These methods, however, require some level of statistical expertise to interpret, or like blur, are difficult to quantify.

VISUALIZING UNCERTAINTY 12 RESEARCH QUESTIONS How can the design of information visualizations commonly found in news media coverage incorporate representations of uncertainty in situations that depict predictive data to facilitate non-expert decision making about current events (specifically economic, political, and weather issues)? How can representations of uncertainty give non-experts a greater appreciation of the possible outcomes associated with a data set? How can interactive representations of uncertainty push non-experts 
 to perform more like experts? How can representations of uncertainty help non-experts quantify uncertainty?

VISUALIZING UNCERTAINTY 13 RESEARCH METHODS Mini Visual Studies Track III Observational Study Editorial Studies Scenarios

VISUALIZING UNCERTAINTY 14 MINI VISUAL STUDIES: DATA SETS Hurricane Path 
 2016 Presidential 
 Presidential Approval Unemployment TBD Projections Polls Ratings Rates Inference Disagreement Completeness Disagreement Type of Uncertainty Disagreement Completeness Inference Inference Precision Inference Location Comparison Distribution Trends Insight Type Trends Trends Comparisons Skeels, Meredith, et al. “Revealing Uncertainty for Information Visualization.” Börner, Katy. Atlas of Knowledge: Anyone Can Map.

VISUALIZING UNCERTAINTY 15 100 10 mean 8 90 7 9 7 80 8 6 6 70 7 7 0 1 2 5 e r m b 1 9 1 8 0 c e 0 2 8 e 2 n e 0 1 D e u 2 60 n J r 6 4.1 J u b e m 5 e c 4.2 4.2 e D 4.1 3.9 4.0 4.2 4.2 4.3 4.3 4.1 4.1 4 4.3 50 4.7 5 3.7 3.7 3.7 3.6 4 3.6 3.6 4.4 4.4 4.0 4.0 3.4 3.5 4.0 4.5 3.2 3.3 3.2 3.3 3.8 4.5 4.5 4.3 5.1 3.8 3.8 40 4 5.0 5.0 3.9 3 3.9 3.9 3 5.5 5.5 30 3 2 2 20 2 1 1 10 1 0 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 100 10 100 10 100 < mean 90 9 90 9 90 mean 7 > mean 80 8 80 8 80 6 70 7 70 7 70 5 60 6 60 6 60 50 5 50 5 50 4 40 4 40 4 40 3 30 3 30 3 30 2 20 2 20 2 20 10 1 10 1 10 1 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 0 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 10 100 10 10 10 Number of Projections Number of Projections < mean < mean < mean 15 and up 15 and up 9 90 9 9 9 mean mean mean 10-14 10-14 5-9 5-9 7 > mean > mean > mean 0-4 1-4 8 80 8 8 8 Mean 0 Mean 6 7 70 7 7 7 6 60 6 6 6 5 5 50 5 5 5 4 4 40 4 4 4 3 3 30 3 3 3 2 20 2 2 2 2 1 10 1 1 1 1 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 0 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 10 10 10 10 10 Number of Projections Number of Projections Number of Projections Number of Projections Density of shadow 15 and up 15 and up 15 and up correlates with number 15 and up Mean 9 9 9 9 9 10 of projections. 10-14 10-14 10-14 10-14 5-9 5-9 5-9 5-9 Width determined by 1-4 1-4 1-4 1-4 number of projections. 8 8 8 8 8 9 Mean 0 0 0 0 Mean Mean Mean 7 7 7 7 7 8 6 6 6 6 6 7 5 5 5 5 5 6 4 4 4 4 4 5 3 4 3 3 3 3 2 2 2 2 2 3 1 1 1 1 1 2 1 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19 Dec-17 Jun-18 Dec-18 Jun-19 Dec-19

VISUALIZING UNCERTAINTY 16 UNEMPLOYMENT PROJECTIONS BY 78 MAJOR US FIRMS Disagreement Uncertainty Trends Analysis

VISUALIZING UNCERTAINTY 17 MGD TRACK III OBSERVATIONAL STUDY Track III studies looked at 2D, interactive, and animated methods for visualizing uncertainty. Matt Norton

VISUALIZING UNCERTAINTY 18 EDITORIAL STUDIES Editorial studies will focus on pictorial and metaphorical ways of incorporating uncertainty, similar to this visualization from The New York Times. “How to Reduce Shootings”, The New York Times