Apps & EcosystemAppsML Canvas
ML Canvas

Run your models on production data.

Deploy custom models on live production data and feed the results straight back into segmentation and personalization — without moving data out of FirstHive.

Embedded
Jupyter notebooks
Scheduled
or on-demand runs
Auto
attribute write-back
ml_canvas.ipynb
df = fh.profiles(curated=True)
model.fit(df, target="churn")
→ writing churn_score to CDP ✓
0accuracy
Model runv4 · scheduled
🧩
Segment-ready
Derived attributes
Part of the FirstHive App Ecosystem
What it does

Custom models, production data, zero data movement

Give BI and data science teams a curated, governed path to run models directly on production data — and make the results instantly usable by marketing.

📓

Run on production data

An embedded Jupyter notebook with curated data access via the FirstHive UI lets BI teams run models directly on production data — no exports, no copies.

🔁

Write back for activation

Derived attributes are automatically ploughed back into the CDP, ready for segmentation and personalization in FirstHive Campaign Manager.

⏱️

Schedule & automate

Run models at will or schedule recurring runs so scores stay fresh.

🎛️

Tune parameters

Adjust feature weights and model parameters through simple user inputs.

How it works

From notebook to campaign

1
Access curated data

Open an embedded notebook with governed access to production data via the FirstHive UI.

2
Build & run your model

Run models directly on live data, on demand or on a recurring schedule.

3
Write derived attributes back

Outputs are automatically ploughed back into unified profiles in the CDP.

4
Activate in Campaign Manager

Use the new attributes for segmentation and personalization — instantly.

Outcomes

Close the gap between data science and activation

Production
Data access
Automated
Write-back
Scheduled
Model runs
Activation
Ready

Put your models to work on live data

Run, schedule and activate custom models — without ever leaving FirstHive.