---
license: gpl
dataset_name: nigerian_energy_and_utilities_energy_theft_anomalies
pretty_name: Nigerian Energy & Utilities – Energy Theft & Anomalies
size_categories: [10K<n<1M, 1M<n<10M]
task_categories: [time-series-forecasting, tabular-regression, other]
tags: [nigeria, energy, utilities, power, grid, smart-meter, renewables]
language: [en]
created: 2025-10-11
---


    # Nigerian Energy & Utilities – Energy Theft & Anomalies

    Anomaly events (tamper, bypass, reversed energy, spikes, zero consumption) with z-scores and flags.

    - **[category]** Smart Meters & Consumption
    - **[rows]** ~80,000
    - **[formats]** CSV + Parquet (snappy)
    - **[geography]** Nigeria (DisCos, substations, plants)

    ## Schema

    | column | dtype |
|---|---|
| event_id | object |
| meter_id | object |
| disco | object |
| timestamp | object |
| anomaly_type | object |
| expected_kwh | float64 |
| observed_kwh | float64 |
| delta_kwh | float64 |
| zscore | float64 |
| flagged | bool |
| resolution_status | object |
| lat | float64 |
| lon | float64 |

    ## Usage

    ```python
    import pandas as pd
    df = pd.read_parquet('data/nigerian_energy_and_utilities_energy_theft_anomalies/nigerian_energy_and_utilities_energy_theft_anomalies.parquet')
    df.head()
    ```

    ```python
    from datasets import load_dataset
    ds = load_dataset('electricsheepafrica/nigerian_energy_and_utilities_energy_theft_anomalies')
    ds
    ```

    ## Notes

    - Data generated with Nigeria-specific parameters (DisCos, tariff bands, 50 Hz grid)
    - Time-of-use shapes and seasonal/weather effects included where applicable
    - Values are internally consistent (e.g., kWh ~ kW*h; voltage/current ~ power)