Data & historisation
Production data that gets lost or hidden away in silos can't be analysed. We build the data foundation that your OEE analyses, AI models and compliance records need.
The data problem in manufacturing
- Data exists but isn't archived, or only in proprietary formats
- No uniform data structure — every machine has its own format
- Reports have to be put together manually in Excel
- Historical data for trend analysis or ML training is missing
- Compliance requires 5–10 years of data archiving, which isn't guaranteed
What we build
Historian system
A time-series database (InfluxDB, TimescaleDB or OSIsoft PI) for machine values, measuring points and process parameters — with defined retention policies.
Unified data layer
A uniform data model across all sources: OPC UA, SQL, REST APIs, ERP exports. The data is readable for BI tools, dashboards and ML pipelines.
Data quality & validation
Plausibility rules, gap handling and anomaly detection right in the data pipeline — before data flows into analysis.
Archiving & compliance
Long-term archiving with defined retention periods, audit logs and immutable records for GMP, ISO and internal requirements.
Technologies we use
Frequently asked questions
Which historian system do you recommend?
That depends on your requirements and budget. InfluxDB is excellent for time-series data at mid-sized companies. For large plants we also use OSIsoft PI or TimescaleDB.
How long should data be stored?
We define that together, based on regulatory requirements (e.g. GMP: 10 years) and your analysis goals. We then size the storage solution accordingly.
Can we combine data from different systems?
Yes. We build a unified data layer that integrates OPC UA, SQL databases, ERP exports and IIoT streams — readable for BI tools and ML models.
Data lake vs. historian — what's the difference?
A historian is optimised for real-time time-series data (machine values, measuring points). A data lake stores all kinds of unstructured data. We combine both, depending on the use case.