How to build beautiful graphs in Python with Seaborn

All the magic happens when the function is called load_datasetwhich expects the name of the data to load and returns dataframe… All these datasets are available in repositories on Github…

Scatter Plot – Scatter Plot

A scatter plot is a chart that displays points based on two dimensions of a dataset. It is very easy to create a scatterplot in seaborn, as it only requires one line of code.

sns.scatterplot(data=flights_data, x="year", y="passengers")

Easy, isn’t it? Function scatterplot takes in the dataset to be rendered and the columns that will act as the x and y axes.

Line Plot – Line Plot

This graph draws a line that represents the development of continuous or categorical data. This kind of graph is popular and famous and easy to create. As before, we will use the function lineplot with a dataset and columns representing the x and y axes. Will do the rest for us seaborn…

sns.lineplot(data=flights_data, x="year", y="passengers")

Bar Chart – Bar Plot

This is probably the most well-known type of chart, and you guessed it we can build this type of chart using seaborn, just as we did it for line and scatterplots using the function barplot…

sns.barplot(data=flights_data, x="year", y="passengers")

She’s very colorful, I know. Later we will learn how to customize it.

Extending functionality with matplotlib

Seaborn builds on matplotlib, extending its functionality and abstracting complexity. At the same time, seaborn does not lose its power. Any seaborn chart can be customized using functions from the matplotlib library. This mechanic can come in handy in certain cases and allows seaborn to take advantage of matplotlib’s capabilities without having to rewrite all of its functions.

Let’s say you want to plot several charts at the same time using seaborn, in this case you can use the function subplot from matplotlib.

diamonds_data = sns.load_dataset('diamonds')
plt.subplot(1, 2, 1)
sns.countplot(x='carat', data=diamonds_data)
plt.subplot(1, 2, 2)
sns.countplot(x='depth', data=diamonds_data)

Using the function subplot several diagrams can be built on one graph. The function takes three parameters: the first is the number of rows, the second is the number of columns, and the third is the number of charts.

We render one seaborn plot on each plot part, mixing matplotlib and seaborn functionality.

Seaborn and Pandas

We have already said that seaborn loves pandas to the point that all of its functions are built on a structure dataframe… Up to this point we have been using seaborn with preloaded data, but what if we want to render data that has already been loaded with pandas?

drinks_df = pd.read_csv("data/drinks.csv")
sns.barplot(x="country", y="beer_servings", data=drinks_df)

Create beautiful graphs with styles

Seaborn gives you the ability to change the interface of your charts. To do this, we have five styles at our disposal out of the box: darkgrid, whitegrid, dark, white and ticks…

sns.set_style("darkgrid")
sns.lineplot(data = data, x = "year", y = "passengers")

Here’s another example.

sns.set_style("whitegrid")
sns.lineplot(data=flights_data, x="year", y="passengers")

Cool use cases

We’ve learned the basics of seaborn, and now let’s put this knowledge into practice and build several charts with the same dataset. In our case, you are using the data set “tips»Which can be downloaded directly from seaborn.

First, let’s load the dataset.

tips_df = sns.load_dataset('tips')
tips_df.head()

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No.	Sun	Dinner	2
one	10.34	1.66	Male	No.	Sun	Dinner	3
2	21.01	3.50	Male	No.	Sun	Dinner	3
3	23.68	3.31	Male	No.	Sun	Dinner	2
4	24.59	3.61	Female	No.	Sun	Dinner	4

I like to display the first few rows of a dataset to get an idea of ​​the columns and the data itself. I usually use several pandas functions to deal with data problems like values null, or add information to the dataset that may be useful. You can read more about this in pandas manual…

Let’s add another column that will be the percentage of the tip for the entire bill.

tips_df["tip_percentage"] = tips_df["tip"] / tips_df["total_bill"]
tips_df.head()

Now the data looks like this:

	total_bill	tip	sex	smoker	day	time	size	tip_percentage
0	16.99	1.01	Female	No.	Sun	Dinner	2	0.059447
one	10.34	1.66	Male	No.	Sun	Dinner	3	0.160542
2	21.01	3.50	Male	No.	Sun	Dinner	3	0.166587
3	23.68	3.31	Male	No.	Sun	Dinner	2	0.139780
4	24.59	3.61	Female	No.	Sun	Dinner	4	0.146808

And now we will start building graphs.

Tip percentage

Let’s try to figure out the percentage of tip and bill. For this we use the function histplotwhich will generate the histogram.

sns.histplot(tips_df["tip_percentage"], binwidth=0.05)

To make everything read well, we had to set up the property binwidth, but now we can understand and evaluate the data faster. Most customers tip 15 to 20% of the bill, but there are a few cases where tips exceed 70%. These values ​​are called anomalies or outliers and are always worth looking at to see if these values ​​are in error.

I was also wondering if the tip percentage changes with the time of day.

sns.histplot(data=tips_df, x="tip_percentage", binwidth=0.05, hue="time")

This time, we created a chart using the entire dataset, not a single column, and set the property hue per column time… This will use a different color for each time value in the chart and add a legend to it.

Total tip for a specific day of the week

Another interesting metric is the number of tips that staff receive depending on the day of the week.

sns.barplot(data=tips_df, x="day", y="tip", estimator=np.sum)

It seems like Friday is a good day to stay at home.

Influence of table size and day of the week on tip

Sometimes you need to understand how several variables affect the final result. For example, how does the day of the week and table size affect tip percentage?

To build the following diagram, we combine the function pivot from pandas for preprocessing and then draw a heatmap.

pivot = tips_df.pivot_table(
    index=["day"],
    columns=["size"],
    values="tip_percentage",
    aggfunc=np.average)
sns.heatmap(pivot)

Conclusion

Of course, there are many more things we can do with seaborn, you can find out about them in official documentation…

I hope you enjoyed this article as much as I did. Thanks for reading!

---

Learn more about the course “Python Developer. Professional “.

Register for the webinar on the topic “Frameworks and Metaclasses”…