Skip to main content

Predict churn

Goal: for every currently active customer, the probability of no activity in the next 30 days.

1. Load your tables​

import pandas as pd
import relativedb

ds = relativedb.from_dataframes(
    {"users": users, "events": events},
    links=[("events", "user_id", "users")])

2. Write the query​

Define churn in the target; restrict the population to active users in WHERE (past-facing window):

PREDICT COUNT(events.*, 0, 30, days) = 0
FOR EACH users.user_id
WHERE COUNT(events.*, -90, 0, days) > 0

3. Score as of today​

df = ds.predict(query, anchor_time=pd.Timestamp.utcnow().normalize())
df.sort_values("probability", ascending=False).head(20)

Each row is entity_id, probability. Users inactive for 90+ days are excluded by the WHERE clause — they already churned.

Variations​

  • Different definition: COUNT(orders.*, 0, 60, days) = 0 (no purchase) or SUM(usage.minutes, 0, 30, days) < 10 (low usage).
  • Backtest: rerun with a past anchor_time and compare against what actually happened — the engine guarantees the context is point-in-time correct.
  • Real model: pass model_backend=relativedb.RtNativeBackend(schema=ds.schema) (guide).

A complete self-checking version lives at examples/industry/growth_churn.py.