How to Choose a Visualization

Line Chart

Good for:

✔ Continuous data (like time)

Bad for:

✘ Many different categorical values (becomes cluttered)
✘ non-continuous values
✘ parts of a whole

Bar Chart:

Good for:

✔ categorical values
✔ easily showing differences
✔ sorting/ ordering / ranking

Bad for:

✘ continuous values
✘ parts of a whole

Pie Chart

Good for:

✔ 1 to 3 values!!!
✔ Percentages
✔ Parts of a whole

Bad for:

✘ more than 4 values (becomes cluttered)
✘ comparing many different categories
✘ non-complete (not parts of a whole)

100% Bar Chart

Good for:

✔ Many categorical values
✔ Percentages
✔ Parts of a whole

Bad for:
✘ non-complete (not parts of a whole)

Dumbbell Chart

Good for:
✔ categories with a minimum and maximum
✔ comparing 2 categories
✔ showing overlap

Bad for:
✘ comparing totals
✘ showing averages
✘ showing distribution

Scatterplot

Scatterplots use 2 numerical values on the x and y-axes and the dots represent observations or categories. This plot shows the relationship between 2 variables as indicated by the pattern (or lack-thereof) of the dots.

Parallel Coordinates Plot

Each parallel line is a variable that can have different units of measurement for different variables. The chords are an observation or category (make sure to label those too). The height and angles of the "chords: illustrate the relationship between variables.

For representing numbers, use one of the following palettes: 

(Top) sequential data (e.g. 1, 2, 3, 4, 5). Change the intensity of a hue.

(Bottom) diverging (e.g. -2 °c,  -1 °c, 0 °c, 1 °c, 2 °c ). Use 2 hues with different names that merge in the middle. 

Change the size of a shape to represent a numerical value. Remember to be cautious when using circles. Humans have a hard time seeing the difference between circle sizes.