Matplotlib and Seaborn

Armed with visualization best practices, we can dive into Matplotlib and Seaborn, the two Python libraries oftentimes used with pandas to create beautiful visualizations.

Seaborn is built on top of Matplotlib and can make more complex visualizations with customized color palettes.

Matplotlib might show up more as a tool for EDA, where the visualizations are mainly tools for us as analysts as opposed to visualizations to present to business leaders.

To import Matplotlib, we can use the following common syntax:

import matplotlib.pyplot as plt

To import Seaborn, use the following:

import seaborn as sns

Bar Charts and Column Charts

Matplotlib

For column charts, you will need to use bar() method and for bar charts, you will need to use the barh() method. Both methods have similar parameters so once you master one, it will be simpler to master the other one.

Let’s revisit the plant nursery.

1
2
3
4

plants = ["Lemon", "Lime", "Orange"]
fruit_yield = [89, 43, 12]

plt.bar(plants, fruit_yield)

The above code would create a column chart that has three different citrus trees on the x-axis and their respective fruit yields for the year on the y-axis.

Seaborn

Seaborn is a little sleeker than Matplotlib in that we have one barplot() method with a parameter that allows us to switch between a column and a bar chart. Seaborn also works best with DataFrames.

If we wanted to make the same plot as above, but with Seaborn and switch it to a bar chart, we would need to do the following:

1
2
3
4

data = [["Lemon", 89], ["Lime", 43], ["Orange", 12]]
citrus = pd.DataFrame(data, columns=["plant","fruit_yield"])

sns.barplot(citrus, x="fruit_yield", y="plant", orient="y")

Setting the orient parameter to "y" specifies that we want a bar chart as opposed to a column chart.

Scatterplots

Matplotlib

We can use the scatter() method to put together a scatterplot in Matplotlib and finally see if there is a relationship between the butterfly population and the number of participants in the pollinator planting program.

1
2
3
4

butterflies=[12,34,89]
pollinator_participants=[10,56,102]

plt.scatter(pollinator_participants, butterflies)

Seaborn

Once again, when working with Seaborn, we need to load our data into a DataFrame before we can try out the scatterplot() method.

1
2
3
4

data = [[12,10], [34,56], [89,102]]
pollinator_program = pd.DataFrame(data, columns=["butterflies","pollinator_participants"])

sns.scatterplot(data=pollinator_program, x="pollinator_participants", "butterflies")

Histograms

Matplotlib

Let’s use histograms to better understand the butterfly population at the nursery. When we chart a histogram, we need to decide on the bin size. Bins are the columns in a histogram. If someone counted the number of butterflies every week for 8 weeks, then it makes sense for us to have 8 bins.

1
2
3

butterflies=[2,6,17,28,59,112,56,24]

plt.hist(butterflies, bins=8)

Seaborn

We can plot a similar histogram using Seaborn and the histplot() method.

1
2
3
4

data = [[1,2], [2,6], [3,17], [4,28], [5,59], [6,112], [7,56], [8,24]]
butterflies = pd.DataFrame(data, columns=["week", "butterflies"])

sns.hisplot(butterflies, bins=8)

Pie Charts

MatplotLib

Finally, let’s explore how we can use these Python libraries to make pie charts. If we needed to create a visualization breaking down the nursery sales by plant category, we could use the pie() method in Matplotlib.

1
2
3
4

categories = ["annuals", "perenials", "trees", "houseplants"]
sales = [678, 1894, 1234, 768]

plt.pie(sales,labels=categories)

The labels parameter specifies what each slice of the pie chart should be called.

Seaborn

As we mentioned on the previous page, not everyone likes pie charts and Seaborn is no exception. Seaborn does not come with a pie chart method.

As you explore making visualizations in Python, we suggest bookmarking both the Matplotlib documentation and the Seaborn documentation . Both sets of documentation come with lots and lots of examples of how you can properly use the multitude of parameters that come with each of the above methods.