Examples / Data science

Predict churn risk

Train a model on past accounts to predict which of this month's accounts are likely to churn, score them, and flag the risky ones for customer success.

IntermediateAbout 15 minutesSchedule: Every day at 17:00Sample files: past_accounts.csv, this_month.csv

What it teaches

Predict task: train
Predict task: score
Model registry
Quality check on predictions

Try it

In Flowcall, open Learn and choose Examples.
Install Predict churn risk.
Press Run now and open the Results tab.

Learn more

1. A model is a task in the Data science guide
2. How to know the model is any good in the Data science guide

The steps

Load past accounts Move a file
Brings past_accounts.csv into the run.
Train churn model Predict
Load this month's accounts Move a file
Brings this_month.csv into the run.
Score this month Predict
Any accounts at risk? Quality check
```
rows >= 1
allowed: predicted_churned 0
```

What it looks like

A frame from the video Predict churn risk: the Flowcall window with the example open. — Predict churn risk · 1:32. The videos are not on the website yet. Every one of them is inside Flowcall: open **Learn** and press **Watch**.