Python Feature Engineering Cookbook

There's more...

We can also aggregate transactional data using the Featuretools library. To do that, let's import the required libraries and load the dataset, just like we did in step 1 to step 4 of this recipe. To work with Featuretools, we need to transform the dataframe into an entity set. Let's get started:

1. Let's create an entity set and give it a representative name:

es = ft.EntitySet(id="customer_data")

For more details on entity sets, take a look at the Creating features from transactions with Featuretools recipe of this chapter.

2. Let's add the dataframe to the entity set by indicating that the transaction_id is the unique transaction identifier and setting the transaction time as the time index of the entity set:

es.entity_from_dataframe(entity_id='transactions', ...