Urban Mobility Data Platform

Overview

A local-first data platform connecting ingestion, validation, analytical modeling, APIs, and dashboard evidence.

Problem

Mobility data projects need a reproducible path from raw inputs to trustworthy analytical outputs without requiring reviewers to provision cloud infrastructure.

Solution

Built a deterministic sample pipeline with data validation, DuckDB and dbt-style analytical modeling, an API, and a reviewer-facing dashboard.

Technical Decisions

• Local-first defaults avoid cloud accounts and private infrastructure in the review path.
• Deterministic fixtures make pipeline behavior repeatable in development and CI.

Outcome

The repository provides a local review path with pipeline guardrails, analytical outputs, API and frontend surfaces, tests, and CI checks.

Key Features

• Deterministic sample pipeline keeps the reviewer path reproducible.
• Validation and guardrails protect the analytical model inputs.
• API and dashboard surfaces expose modeled data for inspection.

Architecture

1. 01

   Sample mobility data

2. 02

   Validation and ingestion

3. 03

   DuckDB storage

4. 04

   dbt-style SQL models

5. 05

   FastAPI service

6. 06

   React dashboard

7. 07

   CI guardrails

Tech Stack

• Python
• DuckDB
• dbt-style SQL
• FastAPI
• React
• CI
• Data validation

Verification

• Deterministic sample pipeline
• Automated tests and CI
• Repository readiness guardrails

Security & Privacy

• No real private mobility data is required for the default workflow.
• Local artifacts and environment files remain outside version control.

Limitations

• The portfolio links to the repository rather than claiming a public live deployment.
• Sample data demonstrates the pipeline contract, not production scale.

Future Improvements

• Add approved production-scale sources when available.
• Expand data-quality monitoring for larger datasets.