Continuous color palettes

Improving Your Data Visualizations in Python

Nick Strayer

Instructor

list of continuous data types on left, non-continuous on the right

blue_scale = sns.light_palette("steelblue")
sns.palplot(blue_scale)

white to blue color palette

red_scale = sns.dark_palette("orangered")
sns.palplot(red_scale)

black to red palette

color encoded heatmap on left and bar chart on right

Keep it simple

indy_oct = pollution.query("year == 2015 & city == 'Indianapolis'")
blue_scale = sns.light_palette("steelblue", as_cmap = True)
sns.heatmap(indy_oct[['O3']], cmap = blue_scale)

light blue heatmap of o3 levels over time

Keep it simple

indy_oct = pollution.query("year == 2015 & city == 'Indianapolis'")
jet_scale = palette = sns.color_palette('jet', as_cmap = True)
sns.heatmap(indy_oct[['O3']], cmap = jet_scale)

heatmap of o3 pollution values with the colorful jet palette

Be aware of color blindness

Avoid transitions between green and red
Palettes that use intensity are safer

green to red equals sad face, light red to red equals happy face

Encoding neutral values

pal_light = sns.diverging_palette(250, 0)
pal_dark  = sns.diverging_palette(250, 0, center = 'dark')

blue to red diverging palette

heatmap of SO2 deviations encoded with blue to red diverging palette

plt.style.use('seaborn-white')

light_palette = sns.light_palette("orangered")
sns.scatterplot(x = 'CO', y = 'NO2', hue = 'O3', data = lb_2012, 
                palette = light_palette)

chart with white background and orange dots highlighting low values as lighter dots

plt.style.use('dark_background')

dark_palette = sns.dark_palette("orangered")
sns.scatterplot(x = 'CO', y = 'NO2', hue = 'O3', data = lb_2012, 
                palette = dark_palette)

chart with black background and orange dots highlighting low values as darker dots

Let's continue in the exercises

Improving Your Data Visualizations in Python