Using construction waste hauling trucks' GPS data to classify earthwork-related locations: A Chengdu case study

TL;DR

This work tackles urban dust pollution by identifying and classifying earthwork-related locations (ERLs) across Chengdu from GPS trajectories of over 16,000 construction waste trucks and 58 urban features. It compares four supervised models (LR, RF, GBDT, MLP) and uses SHAP for feature interpretation, finding Random Forest to be the most effective with an accuracy of $0.785$ and AUROC of $0.870$, aided by a compact feature subset. Key contributors include establishing a grid-based ERL representation, integrating geographic, land cover, POI, and transport features, and deploying the model in the Alpha MAPS system to support real-time dust management; SHAP-guided feature selection further enables model simplification to six features without compromising performance. The practical impact is a scalable, interpretable ERL classification framework that supports city authorities in prioritizing inspections and mitigating urban dust pollution at low personnel costs, demonstrated by 16,132 ERL classifications in December 2023 with $77.8egin{smallmatrix}\\end{smallmatrix}$% accuracy verification.

Abstract

Earthwork-related locations (ERLs), such as construction sites, earth dumping ground, and concrete mixing stations, are major sources of urban dust pollution (particulate matters). The effective management of ERLs is crucial and requires timely and efficient tracking of these locations throughout the city. This work aims to identify and classify urban ERLs using GPS trajectory data of over 16,000 construction waste hauling trucks (CWHTs), as well as 58 urban features encompassing geographic, land cover, POI and transport dimensions. We compare several machine learning models and examine the impact of various spatial-temporal features on classification performance using real-world data in Chengdu, China. The results demonstrate that 77.8% classification accuracy can be achieved with a limited number of features. This classification framework was implemented in the Alpha MAPS system in Chengdu, which has successfully identified 724 construction cites/earth dumping ground, 48 concrete mixing stations, and 80 truck parking locations in the city during December 2023, which has enabled local authority to effectively manage urban dust pollution at low personnel costs.

Figures (8)

• Figure 1: Spatial distribution of ERLs, (Day shift June 5, 2023). Each ERL is represented as a group of small grids of size 200m$\times$200m
• Figure 2: Number of ERLs with CWHTs operation during each shift. (Every 12-hr period is defined as one shift and the dates are from 1 May to 31 June 2023.)
• Figure 3: Raw data used to generate various features.
• Figure 4: Feature importance ranking for different categories of samples.
• Figure 5: Feature importance analysis with all samples for three categories.