Making your visualizations efficient

Improving Your Data Visualizations in Python

Nick Strayer

Instructor

What is efficient?

Reduce the effort needed to see story
Re-organize plots to keep focus
Improve 'ink' to info ratio
Don't compromise the message

Chaotic line simplified over multiple steps to a straight line

An abstract report with three separate plots on the left with arrow point to right showing the same three plots grouped together at the top of the report

# Create a subplot w/ one row & two columns.
f, (ax1, ax2) = plt.subplots(1, 2)

# Pass each axes to respective plot
sns.lineplot('month', 'NO2', 'year', ax=ax1, data=pol_by_month)

sns.barplot('year', 'count', ax=ax2, data=obs_by_year)

A line plot of pollution values over various years next to a bar plot of the number of observations for each year

Clear unnecessary legends

A line plot of pollution values over various years next to a bar plot of the number of observations for each year pointing out that bar plot could be used as legend for line chart

sns.lineplot('month', 'NO2', 'year', ax=ax1, data=pol_by_month, palette='RdBu',)

sns.barplot('year', 'count', 'year', ax=ax2, data=obs_by_year,
            palette='RdBu', dodge=False)

# Remove legends for both plots
ax1.legend_.remove()
ax2.legend_.remove()

A line plot of pollution values over various years next to a bar plot of the number of observations where colors of bar plots bars serve as a legend for line chart

Let's practice

Improving Your Data Visualizations in Python