# Sensor Data and Historian Systems Industrial facilities generate massive time-series data from thousands of sensors: temperature, pressure, flow rate, vibration, composition. This data lives in historian systems, purpose-built databases optimized for high-frequency time-series storage. The big three: OSIsoft PI (now part of AVEVA, dominant in process industries), Aspen InfoPlus.21 (strong in refining/chemicals), and various proprietary SCADA/DCS systems. Each uses different data models, APIs, and naming conventions. This is where bespoke engineering hides. Every plant's historian is configured differently. Sensor tags follow no universal naming convention. Data quality varies wildly: missing values, sensor drift, miscalibrated instruments, manual entries mixed with automated readings. For airports, the equivalent is the Airport Operational Database (AODB), tracking flights, gates, baggage systems, and passenger flow. Different vendor, same integration problem. The implication for any [[Industrial AI MOC]] company: data pipeline construction is the first bottleneck. If you can't automate historian integration, you're sending engineers on-site for weeks per deployment. See [[Bespoke Engineering in Industrial AI]]. Related: [[Knowledge Graphs for Industrial Data]], [[Digital Twins]], [[Bespoke Engineering in Industrial AI]] --- Tags: #deeptech