In order to understand the underlying form of the data, the relationship between the features and response, and more insights, we can use different types of visualization. To understand the relationship between the advertising data features and response, we are going to use a scatterplot.

In order to make different types of visualizations of your data, you can use Matplotlib (https://matplotlib.org/), which is a Python 2D library for making visualizations. To get Matplotlib, you can follow their installation instructions at: https://matplotlib.org/users/installing.html.

Let's import the visualization library Matplotlib:

import matplotlib.pyplot as plt

# The next line will allow us to make inline plots that could appear directly in the notebook
# without poping up in a different window
%matplotlib inline

Now, let's use a scatterplot to visualize the relationship between the advertising data features and response variable:

fig, axs = plt.subplots(1, 3, sharey=True)

# Adding the scatterplots to the grid 
advertising_data.plot(kind='scatter', x='TV', y='sales', ax=axs[0], figsize=(16, 8))
advertising_data.plot(kind='scatter', x='radio', y='sales', ax=axs[1])
advertising_data.plot(kind='scatter', x='newspaper', y='sales', ax=axs[2])

Output:

Now, we need to see how the ads will help increase the sales. So, we need to ask ourselves a couple of questions about that. Worthwhile questions to ask will be something like the relationship between the ads and sales, which kind of ads contribute more to the sales, and the approximate effect of each type of ad on the sales. We will try to answer such questions using a simple linear model.