Seaborn Visuals Quick Guide
Seaborn is one the best statistical visual libraries that Python has to offer. Here is a quick guide that will allow you quick produce visuals easily. For this guide I am using a Avocado data set which you can find below. Check out how to quickly become a data viz master with Data Visualization at Datacamp.
Load Your Dependencies
If you are not familiar with Pandas. Check out our Pandas quick start guide.
Barplots can easily be created by defining the x and y variables and the actual data variable. You can change the orientation of the barplot by add orient ='h'
Line plots are essential and only require the x and y variables. However, you can use a time series plot by adding a datetime value to the x variable.
For more on styling line charts with Seaborn
You can evaluate two continuous variables by using a scatter plot. You can further customize this by using the hue parameter. For more details on styling scatter plots in Seaborn check out this Seaborn scatter plot tutorial.
A modified version of a scatterplot is a regression plot that allows you to have a clear line of best fit which is based on the regression function.
Box plots are great way to evaluate the centrality of your data and find outliers. This follows the typical parameters needed for Seaborn which are defining the X and Y variables and the data source.
A category plot is a modified bar plot that can visualize the raw density of the data underlying data.
A violin plot is the cousin to a box plot. However, the density of the data is distributed along the the y axis.
This plot allows you to to see the frequency of your values distributed in bins. Therefore only a single value is needed to produce this plot.
Join plot allows you to produce a scatter plot and produce a density with marginal histograms.
One of the best ways to see correlation is to use the seaborn heatmap. In order to achieve this you need to produce a correlation of the existing dataframe. You will need to add the corr() function to a data frame.
How to Style a Seaborn Heatmap
A pairplot will show a collective correlation of variables in subplots. This function takes a single variable which is the data frame.
How Create Sublplots in Seaborn
To create a subplots, you are going to eventually chain together your plots by initiating a maplotlib function to set up your subplots. You will need to declare the structure for example the 3,0 parameter points out that there are 3 columns and no rows. you can change this to the structure that best fits your subplots.