This document provides tips to make visualizations more effective by making them simple, easy to understand, and memorable. It recommends using only 2-3 fonts per page, limiting colors and shapes, setting axis ranges carefully, and emphasizing the most important data. Specific tips are given for scatter plots, line plots, area charts, maps, bullet charts, and small multiples. The goal is to present data in a clear, focused manner that guides the viewer's understanding.
2. ✓ Tips to make visualizations more effective
✓ Tips are simple, easy and memorable
✓ The slides cover a variety of visuals (graphs/charts and
techniques)
✓ This document assumes that you already know the basic of
building visualizations.
WHAT WILL WE LEARN?
4. RECOMMENDATION BY “VISUALIZATION WIZARDS”
BEST FONTS
✓ Trebuchet MS or Verdana (especially for tables and numbers)
✓ Arial
✓ Georgia
✓ Tahoma
✓ Times New Roman
✓ Lucida sans
✓ Calibri and Cambria (for tooltips only)
5. MINIMIZE THE DIFFERENCES IN TEXT
✓ Only use 2-3 Fonts/colors per page
✓ Axes and labels should be dark grey (so that it does not distract
viewers)
✓ Make sure the formatting is consistent
✓ For adjacent text, make a change only ONE attribute at a time
(either size, boldness, color).
Picture from Tableau Guidebook
6. COLORS
✓ Tools/Libraries/Packages come with color palettes carefully
designed to work well together, effectively to be applied in many
situations.
✓ Use no more than 2 color palettes.
✓ Do not use more than 12 distinct colors.
✓ There are 3 different types of color palettes: qualitative, sequential,
and diverging. Use the right one for your data.
Example of palettes from Tableau
Picture from Tableau Guidebook
7. SOME OF MY FAVOURITE PALETTES FROM SEABORN (PYTHON)
Qualitative palettes
deep
pastel
dark
Sequential palettes
rocket
mako
GnBu_d
Diverging palettes
BrBG
vlag
reversed vlag
✓ For diverging palettes, the midpoint and endpoints should be meaningful. Zero is often a
meaningful midpoint.
11. ✓ A: The Y axis was set in range 0 - 2900
✓ B: The Y axis was set in range 2500 - 2900
WHY DIFFERENT?
SET YOUR AXIS RANGE CAREFULLY
A B
12. ✓ The human visual system is not very good at
estimating area, especially with slices of circle.
WHY GREAT?
Good visualization Great visualization
PART TO WHOLE
Picture from Tableau Guidebook
13. ✓ Try rotating the view to see which way looks best.
WHY GREAT?
GO VERTICALLY OR HORIZONTALLY
Picture from Tableau Guidebook
14. ✓ Put the most important data on the X and Y axis, less important
data on color, size, or shape.
WHY GREAT?
Good visualization Great visualization
EMPHASIZE THE MOST IMPORTANT DATA
Picture from Tableau Guidebook
15. SCATTER PLOT
✓ Add one more variable to the plot (using colors)
✓ Use transparency to make overlapped data visible
WHY GREAT?
Picture from Tableau Guidebook
16. ✓ Overlay the net profit onto the size of the circles
⇨ Detect the correlation between the 2 factors: it is not the negative correlation between
sales price and quantity, but showing that the company makes the greatest profit on both
ends: high and low sales price.
WHY GREAT?
Good visualization Great visualization
Picture from Tableau Guidebook
17. ✓ The combined chart (B) guides the viewer’s focus toward
a comparison of the 2 trends, showing a clear negative
correlation, as well as the clear net profit information for
decision-making.
DIFFERENT?
A B
Picture from Tableau Guidebook
18. ✓ Adding periods into the visualization (three different
decades) provides different insights into the trend of
the data.
WHY GREAT?
Good visualization Great visualization
LINE PLOT
Picture from Tableau Guidebook
19. ✓ Amplify total funding trends over time.
✓ See how each individual sector contributes to the total
over time.
WHY GREAT?
Good visualization Great visualization
Picture from Tableau Guidebook
20. ✓ The area chart (A) treats each sector as a single pattern.
✓ The bar chart (B) treats each year as a pattern.
DIFFERENT?
A B
Picture from Tableau Guidebook
21. ✓ (A) Break up the data by Triage Acuity category to observe the time the
patients spend in each category.
✓ (B) Break up the data by time segments/bins ⇨ easily observe the peak
treatment length.
✓ (B) Also use colors to observe the distribution in each category.
THESE 2 PLOTS
SHOULD BE USED
TOGETHER
A B
DISTRIBUTION
Triage acuity: the potential severity of a patient's illness or injury.
Picture from Tableau Guidebook
22. ✓ If map cannot be interactive, map is best
when paired with another chart (cross-tab,
bar chart, line chart).
✓ Pie charts are good on maps.
WHY GREAT?
GEOGRAPHICAL DATA
Picture from Tableau Guidebook
23. ✓ Easy to see
✓ Shorten/Make the graph more concise
WHY GREAT?
THE BULLET CHART (BAR CHART WITH REFERENCE LINES)
Picture from Tableau Guidebook
24. BREAK GRAPH TO SMALL MULTIPLES
Not good visualization Great visualization
Picture from Tableau Guidebook
25. ✓ Labels on selection, not all the data point to avoid clutter.
✓ Labels on min/max/the highlighted marks/line ends.
⇨ Can help tell the story quickly and succinctly.
WHY GREAT?
EMPHASIZE THE IMPORTANCES
Picture from Tableau Guidebook
26. It is a long journey to learn how to build
visualization from “good” to “great”.
Be patient and enjoy!