Being able to predict the future can give your business the edge it needs and make it easier to make decisions. In DataChat, you can use the Analyze skill to quickly and easily create machine learning models to help you identify trends and make predictions. In this example, we’ll investigate some customer data to find why customers churn and create a model that can help us predict whether a customer is likely to churn in the future.
Step One: Load Data
The first step is to load our data into a session. Our dataset looks something like this:
With the data loaded, we’re ready to begin investigating.
Step Two: Describe the Data
Before diving into any analysis, we recommend using the Describe skill to get a better understanding of what exactly you’re working with. This can help in your data discovery, and to plan your analysis. The Describe skill gives us some summary statistics about our dataset such as the count of records in a column, the number of unique values in a column, means, minimums, maximums, category types, and display types.
We end up with a table that looks like this:
We can see that there are 21 columns covering everything from the number of dependents a customer has to whether they use our company’s online backup and security products.
Step Three: Analyze Why Customers Churn
Churn analysis helps to identify customer pain points, which then helps your business take steps to improve the customer experience and increase retention rates. In our dataset, the Churn column is most important for this analysis. It indicates whether the customer has churned, or left, our company. We’d like to be able to see what might drive customers to leave and predict whether a current customer will churn in the future. We can use the Analyze skill to do both. Running the Analyze skill on the CHurn column results in an impact chart that looks like this:
We can see that the model has found that the type of contract the customer is using has the biggest impact on whether the customer will churn. The next most impactful features include the customer’s tenure and online security.
From here, we can investigate these findings a bit more with the trend illustrations that DataChat created for us. One of the trends looks like this:
This bubble chart compares each type of contract against customer tenure, grouped into 20-year intervals. From the chart, we can see that it’s specifically month-to-month contracts that lead to the most churn, so we should work to move those customers to one-year or two-year contracts as quickly as possible.
We can save our Analyze model and use it on future datasets to predict whether the customers in that dataset will churn. This helps us to make our analysis more efficient, standardized, and clear.