VisitHGNN: Heterogeneous Graph Neural Networks for Modeling Point-of-Interest Visit Patterns

TL;DR

VisitHGNN addresses the challenge of predicting POI visit distributions with high fidelity by modeling origin-to-destination mobility on a heterogeneous graph that couples POIs and census block groups (CBGs). The method assembles a multi-relational graph with POI–POI, CBG–CBG, and POI–CBG connections, encodes rich POI and CBG features (including BERT text embeddings and ACS socio-demographics), and performs relation-aware fusion before predicting a probability distribution over a distance-based candidate set $N_K(p)$ via a masked KL objective. It demonstrates strong performance on Fulton County mobility data, achieving $D_{KL}=0.287$, $MAE=0.008$, Top-1 $=0.853$, and $R^2=0.892$, substantially outperforming a pairwise MLP and a distance-only baseline, and delivering high $NDCG@50$ and Recall@$k$ values. The work provides calibrated, decision-ready origin maps suitable for urban planning, transportation policy, and public health, while highlighting avenues for inductive generalization, bias auditing, and scalable inference in real-world deployments.

Abstract

Understanding how urban residents travel between neighborhoods and destinations is critical for transportation planning, mobility management, and public health. By mining historical origin-to-destination flow patterns with spatial, temporal, and functional relations among urban places, we estimate probabilities of visits from neighborhoods to specific destinations. These probabilities capture neighborhood-level contributions to citywide vehicular and foot traffic, supporting demand estimation, accessibility assessment, and multimodal planning. Particularly, we introduce VisitHGNN, a heterogeneous, relation-specific graph neural network designed to predict visit probabilities at individual Points of interest (POIs). POIs are characterized using numerical, JSON-derived, and textual attributes, augmented with fixed summaries of POI--POI spatial proximity, temporal co-activity, and brand affinity, while census block groups (CBGs) are described with 72 socio-demographic variables. CBGs are connected via spatial adjacency, and POIs and CBGs are linked through distance-annotated cross-type edges. Inference is constrained to a distance-based candidate set of plausible origin CBGs, and training minimizes a masked Kullback-Leibler (KL) divergence to yield probability distribution across the candidate set. Using weekly mobility data from Fulton County, Georgia, USA, VisitHGNN achieves strong predictive performance with mean KL divergence of 0.287, MAE of 0.008, Top-1 accuracy of 0.853, and R-square of 0.892, substantially outperforming pairwise MLP and distance-only baselines, and aligning closely with empirical visitation patterns (NDCG@50 = 0.966); Recall@5 = 0.611). The resulting distributions closely mirror observed travel behavior with high fidelity, highlighting the model's potential for decision support in urban planning, transportation policy, mobility system design, and public health.

Figures (9)

• Figure 1: High-level pipeline of VisitHGNN
• Figure 2: Heterogeneous graph overview and transductive data split.
• Figure 3: Overview of VisitHGNN
• Figure 4: Prediction and evaluation pipeline in VisitHGNN
• Figure 5: Training dynamics of VisitHGNN.