PostgreSQL • RStudio • Manufacturing Analytics

Manufacturing Asset Health Analytics

A predictive-maintenance analytics project that uses rolling statistical baselines, anomaly thresholds, and sequential event analysis to identify abnormal equipment behavior before failure.

Machine risk summary showing flagged manufacturing sensor readings

Turning IoT readings into equipment-health signals

The analysis evaluates temperature, vibration, voltage, and cycle-time readings from robotic arms, CNC machines, and conveyor equipment. PostgreSQL performs rolling-window feature engineering and anomaly detection, while RStudio produces machine-level summaries and visualization outputs.

360 IoT sensor readings
3 Manufacturing assets
4 Sensor metrics analyzed
20 Readings per rolling window

PostgreSQL data model

DBeaver database structure for the asset health analytics project

Simulated IoT equipment readings

Manufacturing IoT sensor readings displayed in DBeaver

Rolling baselines and z-score alerts

Each reading is compared with the machine's previous operating window. Temperature, vibration, cycle time, and voltage deviations are converted into standardized scores and assigned a health severity.

PostgreSQL anomaly detection results with equipment health status

Sequential event dependency

PostgreSQL LAG functions compare current events with prior voltage and mechanical readings to identify possible event chains such as a voltage drop followed by elevated heat or vibration.

Sequential root cause event analysis performed with PostgreSQL

Equipment-health visualizations

Temperature anomaly chart

Temperature anomalies

Highlights abnormal heat patterns that may indicate mechanical friction, cooling issues, or equipment overload.

Vibration anomaly chart

Vibration anomalies

Surfaces possible bearing wear, alignment problems, or unstable mechanical motion.

Cycle time anomaly chart

Cycle-time delays

Detects production delays that may indicate machine degradation, workflow constraints, or bottlenecks.

Comparison of manufacturing asset alert types

Alert comparison

Compares equipment-health alert categories across the monitored manufacturing assets.

Predictive maintenance through SQL and R

This project demonstrates how manufacturing IoT data can be transformed into equipment-health indicators using PostgreSQL window functions, statistical anomaly thresholds, event sequencing, and R-based data visualization.

Explore the Repository