Flow
Start by taking a look at the flow to see the different steps of data preparation and machine learning that are needed for this project.
EXPLORE !We want to try and predict how much net profit each of our new customers is going to bring us all along our future relationship. In other words, we’re looking to predict our Customer Lifetime Value! A more advanced project for predicting Customer Lifetime Value can be discovered here.
Business Goals
Build a model to predict how valuable a customer will be, based on his first interactions with our website. This way we can make predictions To do this, we’ll be using all types of data: web traffic data, and historical customer purchases, enriched with open data on countries (GDP).
How Do We Do This?
We have several collected different types of data from our historical customers:
- Demographic information (age, gender, location)
- Their actions on our website
- The marketing campaigns they received
- Their previous purchases
In addition, we’re going to use data about their location to enrich our dataset.
Using a preparation step, we’re going to start by joining the different data sources. Then we’ll enrich that data in order to train the model. Then, we can copy the recipe and apply it to new data, so we can score new users using our model!
Explore This Sample Project
Visual Preparation
Take a look at how we clean and analyse historical data, and enrich it with geographic data.
EXPLORE !
Model
Check out our random forest algorithm to predict how much revenue will come from each client.
EXPLORE !
Output Dataset
Finally, check out the output dataset we got after we've deployed the model on our new users. You can see the column predicting the lifetime value on the right.
EXPLORE !