Introduction

Seaborn is a Python library used to create data visualizations (charts) in a simple and aesthetically pleasing way.

This library can typically be used to analyze health data. Here is an example in which respondents were asked, over the previous 30 days, how many days they considered themselves to have been in poor physical health, and likewise for their mental health, with color indicating whether or not they are smokers.

Here is another example showing blood glucose levels according to body mass index (BMI), with people suffering from diabetes shown in black.

Seaborn can be used with different types of data, including Python lists, NumPy arrays, and pandas DataFrames, although pandas DataFrames are generally preferred.

There are different formats for data tables:

Wide-format : 

The wide format is often more natural for humans, and some algorithms or functions expect data in this format. Here is an example:

Long-format : 

The standard format is the long format, which allows data points to be represented using multiple different variables. Here is the previous example in long format:

Here is a description of a table in this format:

It can be useful to check whether any data is missing, particularly if certain algorithms cannot handle missing values or if they could bias the results. Having complete data can also make tasks such as data aggregation easier. Here is the code used to check for missing data:

data=sns.load_dataset("penguins")
print(data.isnull())#on the whole table
print(data.isnull().any())#on each column

Then, if we do not want to consider observations with missing values for a given variable, for example:

data.dropna(subset=["body_mass_g"])

However, caution should be exercised when working with a relatively small dataset. For example, if the missing values consistently occur in the same variable, this may introduce bias. Ultimately, the decision depends on the user's judgment.

penguins.csv

tips.csv

iris.csv

healthexp.csv

glue.csv

diamonds.csv