MLops : Session - 13

Today I'm going to discuss about Data Visualization.

Data Visualization:

Normally we human beings are not good in calculations so try to analyze our dataset using Graphs.

Data visualization is the representation of data or information in a graph, chart, or other visual formats. It communicates the relationships of the data with images. This is important because it allows trends and patterns to be more easily seen.

It is important for both data analysis and data analytics. We use graph also for feature selection. That's why it is so necessary because we can easily understand the graph and do feature elimination manually.

Python has many libraries which we used to draw graphs in which one of the basic is Matplotlib. We also use seaborn (which is an extension of matplotlib) for some more high-quality graphs.

Today we are creating multiple types of graphs using the seaborn library. Let's check out these graphs:

Seaborn internally provide some dataset, I'm using these datasets for creating all the graphs.

Histogram;

Let's check how to create a histogram of a univariate variable.

You can easily tell now by seeing the histogram.

Joint Plot:

It is a type of bivariate graph(involving or depending on two variates). Let's check out some examples of this.

By changing kind parameter, we are able to draw different types of graphs.

Pair Plot:

To plot multiple pairwise bivariate distributions in a dataset, you can use the pairplot() function. 

We also use some parameter inside pairplot() function.

HeatMap:

This graph is very useful for check the correlation for feature selection. We can easily find that which feature is useful or not.

By seeing the heatmap,you easily get that which feature is more correlated or not.

We have many more seaborn plots to draw graphs like barplot, countplot, boxplot, violinplot, stripplot and swarnplot. Besides these, we have some more advanced graphs also like clustermap, implot, etc. Go and check out these on official site.

We are going to see some more usecases of these in our next sessions...

Hope we meet soon...