| --- |
| license: cc-by-nc-4.0 |
| task_categories: |
| - tabular-classification |
| - time-series-forecasting |
| tags: |
| - synthetic |
| - manufacturing |
| - quality-control |
| - iot |
| - sensor-data |
| - predictive-maintenance |
| size_categories: |
| - 10M<n<100M |
| --- |
| |
| # Free Synthetic Manufacturing QC Sensor Readings (100M Rows) |
|
|
| 100,000,000 synthetic sensor readings from a simulated manufacturing environment: 2,000 machines across 10 production lines, each reporting temperature, vibration, pressure, current draw, and humidity readings alongside pass/fail QC outcomes. Built for anomaly detection, predictive maintenance, and quality-control modeling work. |
|
|
| The data isn't uniform noise. Each machine has its own degradation curve between maintenance events, so readings drift more on machines that are overdue for service. QC failures are driven by that drift — anomalous readings fail at a meaningfully higher rate than normal ones, not a flat random rate — so the relationships in the data are ones a model can actually learn. |
|
|
| ## Schema |
|
|
| | Column | Type | Description | |
| |---|---|---| |
| | reading_id | string | Unique reading identifier | |
| | machine_id | string | Machine identifier (MC-0001 to MC-2000) | |
| | production_line | string | Production line (Line-A to Line-J) | |
| | station_id | int | Station number within the line | |
| | timestamp | timestamp | Time of reading | |
| | sensor_type | string | temperature, vibration, pressure, current_draw, humidity | |
| | sensor_value | float | Reading value, in `unit` | |
| | unit | string | Unit for sensor_value | |
| | deviation_pct | float | % deviation from nominal value for that sensor type | |
| | is_anomaly | bool | True if deviation exceeds the anomaly threshold | |
| | qc_result | string | pass or fail | |
| | defect_code | string | Defect code when qc_result is fail; empty otherwise | |
| | shift | string | Day, Evening, or Night | |
| | machine_health_index | float | 0-100 synthetic health score at time of reading | |
| |
| ## Format |
| |
| Single Parquet file, Snappy-compressed. |
| |
| ## Quick start |
| |
| ```python |
| import pandas as pd |
| df = pd.read_parquet("manufacturing_qc_readings_100M.parquet") |
| |
| import datasets |
| ds = datasets.load_dataset("ziadatalabs/FreeSyntheticManufacturingQC100M") |
|
|
| import duckdb |
| duckdb.sql("SELECT * FROM 'manufacturing_qc_readings_100M.parquet' LIMIT 10").show() |
| ``` |
| |
| ## Notes |
| |
| All data is synthetic — no real factory, machine, or production data of any kind. Generated for research, benchmarking, and prototyping use. |
| |
| ## License & Usage |
| |
| CC-BY-NC-4.0. Free for research, learning, and non-commercial use. |
| |
| --- |
| |
| Want more free datasets? Hit the ❤️ and follow. And we take requests — tell us what synthetic data you need, and we'll build it. |
| |
| Zia Data Labs · zia.data.team@protonmail.com |
| |
