Use case

Production output forecasting for capacity and supply planning

Forecast weekly production output from a CSV, use the 80% range to plan capacity, and separate planned downtime from unexpected losses.

Run the sample Put this in production

Production forecasts sit between sales and operations. Sales wants to know what will be available; operations wants to know what will be demanded. The same CSV can answer both if the unit and the calendar are explicit.

The sample is weekly finished units with a slow upward trend. If your plant has planned maintenance weeks, mark them as missing or model the post-maintenance ramp separately; a planned shutdown is not a demand signal.

The sample series

The studio below is preloaded with this weekly finished units series and a 8-step horizon. It is an illustrative example, not a real company’s data. Replace it with your own CSV to run the same workflow.

Show the CSV

week,units
2025-10-06,1240
2025-10-13,1280
2025-10-20,1260
2025-10-27,1310
2025-11-03,1290
2025-11-10,1340
2025-11-17,1320
2025-11-24,1180
2025-12-01,1360
2025-12-08,1390
2025-12-15,1410
2025-12-22,1200
2025-12-29,1150
2026-01-05,1380
2026-01-12,1420
2026-01-19,1440
2026-01-26,1460
2026-02-02,1430
2026-02-09,1480
2026-02-16,1500

CSV forecast studio

Forecast your production output forecasting series.

This studio is preloaded with the weekly finished units example above. Replace it with a CSV of your own history, choose a horizon, and run the forecast. The public demo is free while the Hugging Face Space is available; the API is the paid product.

1 · Paste or upload a CSV2 · Pick a horizon3 · Download the 21 quantiles

Steps

1 credit per API call · the web demo is free while the public Space is available

What to watch in the output

Planned downtime is not a demand drop

A maintenance week should be marked missing or excluded, not treated as a week when customers stopped buying. Otherwise the forecast will carry the shutdown pattern into a normal week.

Forecast the constraint you can act on

If the line is capacity-constrained, output follows capacity, not demand. Forecast demand separately and compare it with the output plan.

Keep quality rejects out of the unit

Finished-good units and started units are different series. Use the unit the downstream planner consumes.

Common mistakes

—Mixing shifts or plants in one column. The aggregate hides the bottleneck.
—Ignoring a supplier constraint that caps output for several weeks. That is a structural break, not noise.
—Using a forecast as a commitment. The band is a planning range; a customer promise needs its own safety margin.

How to put it to work

1. Start with the sample. Run the preloaded series once so you see the output shape before you debug your own CSV.
2. Replace it with your history. Keep one timestamp column and one or more numeric columns. Mark outages and stockouts as empty cells, not zeros.
3. Backtest before you plan. Hold out the last few periods, compare the forecast with what happened, and check whether the actuals land inside the 80% band.
4. Move it into production. Use the REST API for a nightly job or an internal tool. One credit per successful call; failed calls are refunded.

FAQ

Can I forecast multiple production lines?

Yes, up to eight related lines in one file. If the lines share labor or materials, a multivariate forecast can capture the shared pattern; validate it against separate forecasts.

Does the model handle a four-day week?

It learns the weekly pattern from the dates you provide. If the calendar changes, a historical four-day week is not a permanent pattern; mark unusual weeks as missing or add a calendar covariate.

How do I measure forecast quality?

Backtest on held-out weeks and compare MASE, RMSE and 80% coverage. A forecast with a good median but poor coverage is not ready for planning.

Need this as an API?

The same forecast is available as a metered REST endpoint with JSON and a downloadable quantile CSV.

See the API