The road-network-predictability from uclwilson

road-network-predictability's Introduction

Quantifying the Spatial Homogeneity of Urban Road Networks via Graph Neural Networks, Nature Machine Intelligence, 2022.

(Publication DOI: 10.1038/s42256-022-00462-y)

A graph neural network computing the intra-city and inter-city spatial homogeneity of urban road networks (URNs)

Introduction

The spatial homogeneity of URNs measures the similarity of intersection connection styles between the subnetwork and the entire network. It captures the multi-hop node neighborhood relationships, and has promising applications in urban science, network science, and urban computing.
This GitHub repository discusses a user-friendly approach to compute the network homogeneity of URNs worldwide.
URN classification, URN NI calculation, socioeconomic factor relation analysis, inter-city homogeneity analysis are also attached.

Publication

Quantifying Spatial Homogeneity of Urban Road Networks via Graph Neural Networks Jiawei Xue, Nan Jiang, Senwei Liang, Qiyuan Pang, Takahiro Yabe, Satish V Ukkusuri*, Jianzhu Ma*, March, 2022, Nature Machine Intelligence.

Requirements

Window System
Python 3.6
NetworkX 2.1
OSMnx 0.11.4
PyTorch 1.0

Directory Structure

data-collection: Collect and preprocess the road network data for 30 cities in the US, Europe, and Asia.
intra-city-network-homogeneity: Perform the link prediction on URNs using 6 different encoders (such as relational GCN) and 1 decoder (DistMult) and compute F1 scores.
road-classification: Implement the URN classification and discover its connections with F1 scores.
association-analysis: Conduct the correlation analysis between F1 scores and social-economic factors, network topology metrics.
inter-city-network-homogeneity: Get the inter-city homogeneity by learning URN features on city A and testing on city B.