Tequila Industry Wastewater Chemical Oxygen Demand Predictions from IoT Sensors

Tequila, Jalisco, Mexico is the location for this dataset of 4,038 records collected over 87 days from IoT sensors monitoring industrial wastewater. The data includes measurements of suspended solids, dissolved oxygen, turbidity, and electrical conductivity, used to predict chemical oxygen demand (COD) with machine learning models. It was created by Alfredo Figarola-Figarola and last updated on 2026-04-17.

Use Cases

• Train machine learning models to predict chemical oxygen demand based on sensor measurements like suspended solids and dissolved oxygen.
• Benchmark model performance (e.g., random forest, XGBoost, gradient boosting) for water quality prediction tasks.
• Analyze sensor data patterns over an 87-day period to understand wastewater variability in the tequila industry.
• Investigate model limitations in predicting extreme COD values as noted in the residual analysis.

Strengths

• Contains 4,038 records, providing a substantial sample for model development.
• Data was collected over a specific 87-day period, offering a defined temporal scope.
• Model validation results are provided, with gradient boosting achieving an R² of 0.9878.
• Residual analysis indicated homoscedasticity and approximate normality, suggesting model robustness.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is small at 239.7 KB, which may limit the complexity of models it can support.
• The description notes that the evaluated models struggle to predict exceedingly low or high COD values.

Provenance

Source

figshare, authored by Alfredo Figarola-Figarola.

Collection Method

Collected via IoT sensors monitoring wastewater from the tequila industry.

Time Range

Data collected over an 87-day period (specific dates not provided).

Freshness

Last updated 2026-04-17 04:12:07; freshness should be verified.

Geography

Tequila, Jalisco, Mexico.