---
license: other
language:
- en
task_categories:
- tabular-classification
- tabular-regression
size_categories:
- 100K<n<1M
pretty_name: "Africa Synthetic Telecom Hardware Sensor Data Nigeria (TsFile)"
tags:
- africa
- electric-sheep-africa
- open-data
- metadata-backed
- energy
- parquet
- tabular
- text
- telecom
- hardware
- sensor
- data
- power
- tsfile
- modality:timeseries
- timeseries
- format:tsfile
configs:
- config_name: default
  data_files:
  - split: train
    path: staged.tsfile
modality:
- timeseries
- tabular
- text
---

# Africa Synthetic Telecom Hardware Sensor Data Nigeria (TsFile)

This dataset is an Apache TsFile conversion of [`electricsheepafrica/africa-synth-telecom-hardware-sensor-data-nigeria`](https://huggingface.co/datasets/electricsheepafrica/africa-synth-telecom-hardware-sensor-data-nigeria), a synthetic Nigerian telecom tower hardware sensor dataset with temperature, power, voltage, humidity, vibration, health-status, and alert readings.

## Source Dataset

- **Original dataset:** [`electricsheepafrica/africa-synth-telecom-hardware-sensor-data-nigeria`](https://huggingface.co/datasets/electricsheepafrica/africa-synth-telecom-hardware-sensor-data-nigeria)
- **Source files:** `hardware_sensor_data.parquet` and `hardware_sensor_data.csv`
- **Rows:** 500,000
- **Columns:** 12
- **Generated date in source card:** 2025-10-05
- **Time range in converted data:** 2025-09-01 00:00:00 to 2025-10-01 23:59:00
- **License note:** the source card metadata declares `gpl`; the source README body also says "MIT License - For educational and research purposes".

## TsFile Conversion

- **Converted file:** `staged.tsfile`
- **TsFile size:** 690,291,491 bytes
- **TsFile table:** `hardware_sensor_data`
- **Time column:** source `timestamp` is parsed as the TsFile `Time` column with millisecond precision.
- **TAG column:** `sensor_id`; the source has 500,000 unique sensor IDs, so `(sensor_id, Time)` is unique.
- **Rows preserved:** 500,000 source rows -> 500,000 staged rows.
- **Columns preserved:** all 12 source columns are represented; no source columns or rows are dropped.

## Schema

| Role | Column | Type | Notes |
|---|---|---|---|
| TIME | `Time` | INT64 ms | Renamed from source `timestamp` during staging |
| TAG | `sensor_id` | STRING | TsFile device dimension |
| FIELD | `tower_id` | STRING | Tower identifier |
| FIELD | `city` | STRING | Nigerian city |
| FIELD | `equipment_type` | STRING | Hardware category |
| FIELD | `temperature_celsius` | DOUBLE | Sensor measurement |
| FIELD | `power_draw_watts` | DOUBLE | Sensor measurement |
| FIELD | `voltage_v` | DOUBLE | Sensor measurement |
| FIELD | `humidity_percent` | DOUBLE | Sensor measurement |
| FIELD | `vibration_level` | DOUBLE | Sensor measurement |
| FIELD | `health_status` | STRING | Values observed locally: `normal`, `warning` |
| FIELD | `alert_triggered` | BOOLEAN | 20,508 true values in the converted data |

## Read Example

Use Apache TsFile tooling or SDKs to read:

```python
from tsfile import TsFileReader

path = "staged.tsfile"
reader = TsFileReader(path)
schemas = reader.get_all_table_schemas()
reader.close()
```