HERMES

Research for Mobility Evaluation and Simulation

Open-source research framework for urban mobility simulation Urban Data Science • Sustainable Mobility • Energy Transition

---

Overview

HERMES is an open-source spatial simulation framework for modeling bicycle adoption and assessing its potential mobility, energy and climate impacts.

The project integrates demographic, socioeconomic, mobility, geospatial and environmental data to represent the territorial factors that influence the feasibility and adoption of cycling.

Its first case study focuses on Villefranche-sur-Saône and its surrounding mobility system, where HERMES is being developed to simulate bicycle and electric bicycle adoption under alternative scenarios.

Rather than producing a single predictive model, HERMES combines reproducible data engineering, spatial modelling and scenario simulation to explore how changes in mobility behaviour could affect transport patterns, energy demand and environmental impacts.

---

Current Project Status

HERMES is currently in the territorial data engineering and feature construction phase.

The implemented components focus on collecting, preprocessing and integrating the demographic, employment, commuting, geographic, climate and topographic data required by the future simulation engine.

The cycling feasibility models, behavioural adoption models, economic impact calculations and scenario simulation components are under development.

---

Quick Start

Clone the repository

git clone https://github.com/rsquaredata/hermes.git
cd hermes

Create the environment

conda env create -f environment.yml
conda activate hermes

Install the package

pip install -e .

Launch Jupyter Lab

jupyter lab

Open the notebooks in the notebooks/ directory to reproduce the complete data preparation workflow.

The current pipeline automatically:

• downloads public datasets
• preprocesses raw data
• builds standardized municipality-level tables
• exports integrated datasets in Parquet format

---

Data Pipeline

flowchart TD

    A[Open Data Sources]

    A --> B[Download]
    B --> C[Loaders]
    C --> D[Preprocessing]

    D --> E[Population]
    D --> F[Employment]
    D --> G[Workplace Employment]
    D --> H[Mobility Flows]
    D --> I[Municipality Boundaries]
    D --> J[Topography]
    D --> L[Climate]

    E --> K[Municipality Dataset]
    F --> K
    G --> K
    I --> K
    J --> K

    H --> R[Mobility Network]

    K --> S[Feature Engineering]

    S --> T[Behaviour Models]
    R --> T

    T --> U[Scenario Simulation]
    U --> V[Impact Assessment]
    V --> W[Decision Support]

Loading

---

Current Datasets

The current data preparation pipeline includes:

Dataset	Status
Population	✅
Employment	✅
Workplace Employment	✅
Mobility Flows	✅
Municipality Boundaries	✅
Topography	✅
Climate (SAFRAN/SIM)	✅

---

Current Outputs

The preprocessing pipeline currently produces the following standardized datasets:

• population.parquet
• employment.parquet
• workplace_employment.parquet
• mobility.parquet
• municipality_boundaries.parquet
• topography.parquet
• climate.parquet

These datasets constitute the territorial data layer used by the HERMES simulation engine.

---

Initial Case Study

The first HERMES case study focuses on Villefranche-sur-Saône (France) and its surrounding mobility system.

The objective is to model the potential adoption of conventional and electric bicycles by accounting for territorial constraints and individual mobility conditions, including commuting patterns, distance, topography, climate and socioeconomic characteristics.

Alternative adoption scenarios will then be used to estimate their potential effects on mobility patterns, transport energy demand and climate-related impacts.

Villefranche-sur-Saône serves as the initial experimental territory for developing and validating the HERMES methodology, with the longer-term goal of making the framework transferable to other territories.

---

Research Questions

HERMES aims to investigate the following questions:

1. Cycling feasibility
   For a given commuting flow, is cycling or electric cycling technically feasible when accounting for distance, elevation gain, slope, climate and available infrastructure?

2. Individual economic impact
   How much could a commuter save by replacing car journeys with a conventional or electric bicycle, considering energy, fuel, maintenance, parking and equipment costs?

3. Large-scale modal shift
   What would happen if a significant proportion of commuters shifted from private cars to bicycles or electric bicycles?

4. Territorial impacts
   How would alternative cycling adoption scenarios affect transport energy demand, greenhouse gas emissions, traffic patterns and inequalities between municipalities?

5. Policy evaluation
   Which combinations of cycling infrastructure, electric bicycle subsidies and mobility policies could produce the largest feasible modal shift?

---

Repository Structure

HERMES/
├── data/
│   ├── raw/
│   ├── prepared/
│   ├── features/
│   ├── dimensions/
│   ├── scenarios/
│   └── external/
│
├── notebooks/
│   ├── 01_population.ipynb
│   ├── 01bis_mayotte_population.ipynb
│   ├── 02_employment.ipynb
│   ├── 02bis_mayotte_employment.ipynb
│   ├── 03_workplace_employment.ipynb
│   ├── 03bis_mayotte_workplace_employment.ipynb
│   ├── 04_mobility.ipynb
│   ├── 05_municipality_boundaries.ipynb
│   ├── 06_topography.ipynb
│   ├── 07_climate.ipynb
│   ├── 08_municipality_table.ipynb
│   ├── 09_territorial_graph.ipynb
│   ├── 10_graph_exploration.ipynb
│   ├── 11_network_analysis.ipynb
│   └── 12_terrain_features.ipynb
│
├── src/
│   └── hermes/
│       ├── preprocessing/
│       ├── integration/
│       ├── features/         # planned feature engineering modules
│       ├── scenarios/        # planned scenario definitions
│       ├── simulation/       # simulation engine under development
│       ├── raster/
│       ├── terrain/          # terrain analysis under development
│       ├── agents/           # planned decision-support agents
│       ├── sources/
│       └── ...
│
├── tests/
└── README.md

---

Roadmap

Territorial Data Layer

• Population and socioeconomic data
• Employment and commuting flows
• Municipality boundaries
• Climate data
• High-resolution elevation data
• Terrain and cycling-relevant slope indicators
• Cycling infrastructure
• Land use and accessibility indicators

Mobility Modelling

• Origin-destination mobility representation
• Cycling feasibility indicators
• Bicycle adoption model
• Electric bicycle adoption model
• Behavioural and territorial constraints

Scenario Simulation

• Baseline mobility scenario
• Conventional bicycle adoption scenarios
• Electric bicycle adoption scenarios
• Progressive adoption levels
• Large-scale modal shift scenarios
• Cycling infrastructure expansion scenario
• Electric bicycle subsidy scenario
• Combined policy scenarios
• Sensitivity and uncertainty analysis
• Spatial comparison of scenario outcomes

Impact Assessment

• Modal shift
• User mobility costs and potential savings
• Bicycle and electric bicycle ownership costs
• Transport energy demand
• Greenhouse gas emissions
• Traffic and congestion impacts
• Distributional and territorial impacts

Decision Support

• Scenario comparison
• Spatial visualisation
• Interactive exploration tools

---

Project Philosophy

HERMES follows a modular and reproducible design philosophy.

The project separates data acquisition, preprocessing, integration, modelling and simulation into independent components to facilitate reuse, transparency and future extensions.

All datasets originate from publicly available sources and can be automatically reproduced through the data pipeline.

---

License

This project is released under the MIT License.

See the LICENSE file for details.

---

Disclaimer

This project is intended for research, educational, and portfolio purposes.

Simulation results should be interpreted as decision-support estimates rather than exact forecasts and depend on the assumptions and quality of the underlying data.

HERMES — Research for Mobility Evaluation and Simulation