Concepts and Strategies for Developing Effective Data Visualizations Becky Bates GCE Solutions 1
Outline • Key Drivers for Effective Visuals • Design Considerations • Do’s and Don’ts • Putting it all together 2
Key Drivers of Effective Data Visuals 3
Telling The Story Which graph is better? Figure B Figure A Iwurk Biomarker ug/dl Iwurk Biomarker ug/dl Time Time Picme Nogude 4
Information Transfer Rate * ITR = time it takes for a individual to understand the information (i.e. story) being depicted in the data visual Goal it to minimize this time as much as possible 5 *Edward Tufte concept
Venue Visualization considerations greatly depend on the venue: Presentation Poster Journal Regulatory Agency Internal Scientific Meetings Internal Executive Meetings 6
Design Considerations 7
Type of Graph Bar Charts Bubble Plot Pie Chart Line Graphs Heat Map Scatter Plot Box Plot Tree Map 8
Graphing Elements Details Movement Fonts Color Ordering and Juxta positioning Lines 9
Visual Comparisons: Shape Matters Which set is Blue twice the size of Yellow? Set 1 Set 2 Set 3 10
Visualization Do’s and Don’ts 11
Visualization Litmus Test Does the visual maintain the mathematical relationship? Yes – your on your way No – stop and fix 12
Preservation of Comparison Don’t Do Figure B 100 Figure A 75 100 50 75 % 50 25 25 0 Blue looks 2 x Larger than Orange Blue is actually is 0.5 x Larger than Orange Key concept with bar charts : Length of bar informs the reader of the magnitude. Need to preserve that relationship. This is especially important when comparing 2 groups.
Axes Consistency Do Don’t 100 100 100 25 75 75 75 20 % 50 50 50 10 25 25 5 25 0 0 0 0 Mortality Rate SAE Bleed Rate Mortality Rate SAE Bleed Rate Bleed rates look similar to mortality rates Bleed rates are actually less than the mortality rate Key concept with consistency : When creating visualizations across a presentation, need to keep consistent scales throughout for similar data especially if comparisons between the measures will be part of the story.
Preservation of Time Scale Don’t Do Figure A Figure B 6 10 0 0 2 10 2 4 Key concept with line graphs The slope of the line conveys to the reader the rate of increase/decrease. This slope is preserved only when the numerical scale is preserved.
Avoid Chart Junk Chart Junk : Anything that does not contribute to or distracts the user’s understanding of the “data story”
Putting it all together 17
Trial Example Therapeutic Area: Nomophobia Fear of being without mobile phone coverage (http://www.telegraph.co.uk/news/science/10289366/13-of-the-most-unusual-phobias.html) Trial: D ouble-blind I nternational S tudy C omparing O ral N omophone vs N othing in E xasperated C ellular T elephone users DISCONNECT Treatments: 150 mg Nomophone vs Nothing Baseline Measures: Age (yrs) Age 1 st Cell Phone Age 1 st Cell Phone Class (<13 yrs, 13-18 yrs, 19-24 yrs, >=25 yrs) Length of Cell Exposure (Age – Age 1 st Cell) Nomophobia Questionaire (NMP-Q)
DISCONNECT Primary Endpoint Primary Measure: I nternational C ellular A ssessment of N omophobia T olerance Scale ICANT Scale Definition: Take phone away until patient demands to have it back. Earn 1 pt every 15 mins without phone Earn 1 pt for each of the following every 15 mins without phone: Change in heart rate < 5 bpm Figgity Scale <= 2 Lack of Focus Scale <= 2 Anxiety Scale <= 2 ICANT Score = sum of all points earned while phone was removed Primary Enpoint: % Responders Responder = ICANT Scale > 20 pts
Color and Movement Example NMP-Q Score
Color Example
Lines, Font, and Scale Example
Lines, Font, and Scale Example
Spatial vs Temporal and Order Example Responder Visual
Spatial vs Temporal and Order Example Sub-score Visual
Spatial vs Temporal and Order Example Responder and Subscore Visual Combined
In Summary Key Drivers What is your story How quicky is your story being conveyed Venue Elements to considers: Graph type Spatial Arrangement Ordering of information Colors Lines Fonts Movement Details
THANK YOU 28
Recommend
More recommend
