Estimation of on-road mobile emissions using machine learning: A case study in Panama City

This dataset provides a high-resolution inventory of on-road vehicle emissions in Panama City, Panama. The data was synthesized using a novel framework that combines computer vision for real-time vehicle classification and Tier 2 emission factor modeling. It captures the spatiotemporal dynamics of air pollutants across various critical traffic points in the metropolitan area. Data Content The primary file, emissions_dataset.csv, contains individual vehicle detection records and their corresponding estimated emissions based on vehicle type, fuel type, and local traffic conditions. File Structure: emissions_dataset.csv vehicle: Unique identifier for each detection, including the timestamp of the image processed. classifier-tier1: Broad vehicle category (e.g., PC: Passenger Car, LCV: Light Commercial Vehicle, HDV: Heavy Duty Vehicle, L-CAT: L-Category/Motorcycles). classifier-PC: Specific sub-category for Passenger Cars (e.g., SUV, HS: Hatchback/Sedan, TAXI). Fuel: Type of fuel used by the vehicle (Petrol / Diesel). Fuel_Consumption: Estimated fuel consumption for the specific vehicle type (kg/km). CO: Carbon Monoxide emissions (g/km). NMVOC: Non-Methane Volatile Organic Compounds emissions (g/km). NOx: Nitrogen Oxides (NO and NO2) emissions (g/km). PM: Particulate Matter emissions (g/km). N2O: Nitrous Oxide emissions (g/km). NH3: Ammonia emissions (g/km). IDP: Indeno(1,2,3-cd)pyrene emissions (PAH) (g/km). BKF: Benzo(k)fluoranthene emissions (PAH) (g/km). CO2: Carbon Dioxide emissions (kg/km). SO2: Sulfur Dioxide emissions (g/km). Lugar: Monitoring location within Panama City (e.g., Centenario, Rio Abajo, El Ingenio, Tocumen). Hora: Time block of the observation (e.g., 8AM, 12PM). Methodology Summary Vehicle Detection & Classification: Using Deep Learning models (Computer Vision) to identify and categorize vehicles from traffic footage at strategic locations in Panama City. Emission Factor Assignment: Applying EMEP/EEA methodology adapted to the Panamanian vehicle fleet age and fuel quality standards. Spatiotemporal Analysis: Emissions are aggregated by location (Lugar) and hour (Hora) to identify peak pollution periods and high-emission urban corridors. Usage Notes Missing Values: In classifier-PC, values are only present for the "PC" (Passenger Car) Tier 1 category. Geographic Scope: The data covers major thoroughfares including Vía España, Tumba Muerto, Ave. Balboa, and entry points like Centenario. License This dataset is made available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. Citation If you use this dataset in your research, please cite the original paper: Author(s). (Year). Spatiotemporal estimation of on-road mobile emissions using machine learning: A case study in Panama City. [Journal Name / DOI].