Ecommerce: Should we focus on mobile or web development?

Background

An ecommerce company based in New York City sells clothing online and they also have in-store style and clothing advice sessions.

Technologies

Python, Pandas, Numpy, Seaborn, Scikit-learn

Goal

Based on a customer dataset given by the company, find out whether the company should focus on developing their mobile app or web app to best serve the customers that will , in turn, generate the most profit for the company.

Let's begin!

Step 1: Explore the given dataset

Let's read in the dataset


                import pandas as pd

                customers = pd.read_csv("ecomm_customers")

Check what columns are given


                list(customers.columns)

>>> ['Email', 'Address', 'Avatar', 'Avg. Session Length', 'Time on App', 'Time on Website', 'Length of Membership', 'Yearly Amount Spent']

Since our goal is to maximize the company's profit, "Yearly Amount Spent" seems to be a crucial feature we will be focusing on.
"Time on App" and "Time on Website" also seem to be important to answer the original question "Mobile or Web?".

Now, let's visualize the correlation between each of the features using Seaborn.


                import seaborn as sns

                sns.heatmap(customers.corr(), annot=True)

A correlation value of 1 means that the two features correlate to each other completely. For example, it makes sense that "Time on App" is 100% correlated with "Time on App" itself. We can safely ignore the 1s in this plot.

We can see there is a very high correlation value of 0.81 between "Yearly Amount Spent" and "Length of Membership". Hence, we know that the longer the membership, the more amount that customer will spend.

Since this insight doesn't give a direct answer to our original question of "Mobile or Web?", let's continue exploring the data given.

Visualize "Time on App" vs "Yearly Amount Spent"


                    sns.jointplot(x="Time on App", y="Yearly Amount Spent", data=customers)

"Time on Website" vs "Yearly Amount Spent"


                    sns.jointplot(x="Time on Website", y="Yearly Amount Spent", data=customers)

From the above two plots, we can see "Time on App" has a more linear relationship to "Yearly Amount Spent" (more time on the mobile app, more profit to the company).

At this point, we can already conclude that the mobile app is more profitable than the website.

However, let us still build a Linear Regression model to see if we missed out any crucial components


                    from sklearn.model_selection import train_test_split


                    y = customers['Yearly Amount Spent']

                    X = customers[['Avg. Session Length', 'Time on App', 'Time on Website', 'Length of Membership']]

                    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)

build the model:


                    from sklearn.linear_model import LinearRegression


                    lm = LinearRegression()

                    lm.fit(X_train, y_train)

Investigate our model's accuracy:


                    predictions = lm.predict(X_test)



                    from sklearn import metrics

                    import numpy as np

                    print("RMSE:", np.sqrt(metrics.mean_squared_error(y_test, predictions)))

>>> RMSE: 8.861125833738813

A mean error of only $8.86 for the predictions which I'd say that the model is rather accurate.

Therefore, we can rely on the model to see how all the features relate to company's profit ("Yearly Amount Spent"):


                    coeffs = pd.DataFrame(lm.coef_, X.columns)

                    coeffs.columns = ['Coefficient']

                    coeffs

Interpretation:
1. Every 1 unit (minute) increase in time spent on the website would increase ~$0.499 in company's profit.
2. Every 1 unit (minute) increase in time spent on the mobile app would increase ~$38.325 in company's profit.

Conclusion

The analysis suggests that the mobile app is 76 times more profitable than the website.
However, it is unclear from the data whether the website is an important piece that supports the mobile experience.
Hence, I suggest that the company invest just enough to make the website functional and decent. And focus the rest of the available effort into optimizing the mobile experience and bringing customers onboard to using the mobile app.

The End.
I hope you enjoyed the reading. <3