Data Visualization - Teachers As Scholars Program

Yesterday, I attended the first of two sessions in the Teachers As Scholars program entitled Data Visualization.  The course is being taught by Hanspeter Pfister, a Harvard Computer Science professor.  The topic was very interesting and I thought I would share the five important principles to creating effective graphs (visulaizations) which we learned in class. 

1. Graphical Integrity.  
   Don't distort your data!  Start your bar charts and line graphs at zero and if you do not, make sure you add a disclaimer.  Make sure that your pie chart percentages add up to 100%.
2. Keep it Simple.
   If it doesn't convey data, remove it.  This might include color, lines, shadows or more.
3. Use the right chart for your information type.
              Comparison -- bar charts
              Trends -- line charts
              Correlation -- scatter plot
              Distribution -- histogram
              Proportions -- pie chart
4. Use color strategically.
   
   10% of the male population is color blind and have difficulty distinguishing between red and green.  Take this into consideration when you are creating a graph.  Consider using color as follows:
   If you are ranking items in a category:
             Luminance -- the light to dark of a color
             Saturation -- how far away from gray
             Hue -- Color (e.g., red, green, blue)
   Check out ColorBrewer2.org a tool of color scales to help select colors for graphs.
5. Tell a story with your data.
   This helps your audience to gain insight from the data.  For more information on this, check out How to Tell a Story With Your Data by Jim Strikeleather in the Harvard Business Review.