README.md

metadata

language:
  - pt
license: cc-by-4.0
tags:
  - urban-mobility
  - public-transport
  - graph-theory
  - network-analysis
  - centrality-metrics
  - sao-paulo
  - brazil
  - gtfs
size_categories:
  - 10K<n<100K
task_categories:
  - tabular-classification
  - graph-ml

São Paulo Transit Network — Centrality Metrics

Description

Dataset containing graph centrality metrics for 21,892 bus stops in the São Paulo public transit network, built from official SPTrans GTFS data.

The graph was modeled with stops as nodes and consecutive connections within each trip as edges.

Structure

Column	Type	Description
stop_id	str	Unique stop identifier (SPTrans)
stop_name	str	Stop name/address
lat	float	Latitude
lon	float	Longitude
degree	int	Node degree (number of direct connections)
degree_centrality	float	Normalized degree centrality
betweenness_centrality	float	Betweenness centrality
closeness_centrality	float	Closeness centrality

Graph Statistics

• Nodes: 21,892 stops
• Edges: ~29,797 unique connections
• Source: SPTrans GTFS

Usage

from datasets import load_dataset

ds = load_dataset("cintia-shinoda/sp-transit-network-centrality")
print(ds["train"][0])

Citation

@misc{shinoda2026sp-transit,
  author = {Cintia I. Shinoda},
  title = {São Paulo Transit Network — Centrality Metrics},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/cintia-shinoda/sp-transit-network-centrality}
}

---

Descrição (PT-BR)

Dataset com métricas de centralidade de grafo para 21.892 paradas de ônibus da rede de transporte público de São Paulo, construído a partir do dados GTFS da SPTrans.

O grafo foi modelado com paradas como nós e conexões consecutivas dentro de cada trip como arestas.

Contexto Acadêmico

Preparado como parte do Trabalho de Conclusão de Curso (TCC) em Ciência de Dados na UNIVESP (2026), com foco em análise de redes de transporte urbano usando Teoria dos Grafos.