How much historian data do ML platforms need?

Enough normal production to cover real operating regimes: commonly 6-24 months backfill. More products and seasonality push toward the high end; a stable mono-product line can start lower.

Do we need to clean the data first?

Not exhaustively: modern platforms tolerate gaps and noise. Fix what changes conclusions: stored resolution, tag mapping for the variables that matter, and regime context (product/lot/shift). A perfection-first cleansing project is where momentum goes to die.

Home / Encyclopedia / Historian data (and ML)

Data · Entry · updated September 1, 2026

Historian data (and ML)

Key takeaways

A historian is the plant's time-series archive: years of sensor and PLC data, already collected, already paid for.
It's the fastest path to ML value: modern platforms train on historian data with no new hardware.
The traps are compression artifacts, meaningless tag names, and missing context (product, shift, lot): fixable, but budget for them.
You don't need a data lake first: platforms increasingly connect straight to the historian and contextualize as they learn.

Every ML ambition in a plant runs through one unglamorous asset: the historian. OSIsoft PI, Wonderware, GE Proficy, or a homegrown SQL store: it holds years of the exact data models need, and it's already there. Plants that start ML from the historian get first results in days; plants that start by building a data lake often get their first results in the following fiscal year.

The TEEPTRAK approach

Predictive models need stop history. Most plants do not have it yet.

Before a model can predict a failure, it needs a record of past stops: when, how long, on which machine and why. TEEPTRAK builds that record from the first day, and the same data already cuts unplanned downtime through faster reaction and clear loss priorities.

Every stop captured with its start, end and duration
Reasons entered at the machine, so breakdowns are told apart from changeovers and waiting
History exportable to any maintenance or analytics tool

Used in 450+ factories. At Hutchinson, OEE rose from 47% to 72%. Installed with a module wired into the electrical cabinet and a sensor at the edge of the line. No PLC changes. TEEPTRAK is not an MES: it measures, so you know what to fix first.

See how TEEPTRAK works →Map OEE data to ML features →

What models need from it

Three things, in order of pain. Resolution: historians compress (deadband, swinging-door); aggressive compression can erase the very transients a model needs. Check actual stored resolution against what the failure modes require. Tag semantics: models don't care that a tag is called TT4711, but engineers acting on attributions do, a tag-mapping pass is part of every serious deployment. Context: product codes, lot IDs, shift and changeover markers turn anonymous time-series into regime-aware training data; without them models confuse a product change with a process drift.

Practical sequence

Connect the historian read-only, backfill 6-24 months, map the tags that matter, add context feeds (MES/ERP product and lot data), and let the platform train. Vendors differ enormously in how much of this they automate: it's a better demo question than anything about algorithms.

FAQ

How much historian data do ML platforms need?: Enough normal production to cover real operating regimes: commonly 6-24 months backfill. More products and seasonality push toward the high end; a stable mono-product line can start lower.
Do we need to clean the data first?: Not exhaustively: modern platforms tolerate gaps and noise. Fix what changes conclusions: stored resolution, tag mapping for the variables that matter, and regime context (product/lot/shift). A perfection-first cleansing project is where momentum goes to die.

Related entries

Comparing platforms that do this? The adoption data pack carries the market numbers behind this entry.

From theory to signals

Which machine signals become which features, and how each breaks

Free, no login. One follow-up from TEEPTRAK, then nothing.

Get the mapping kit →