An end-to-end analytics platform that transforms Airbnb listings into a modern data warehouse and interactive dashboard using AWS, dbt, and BI-as-code.
This project builds a complete analytics pipeline for Airbnb listings in São Paulo.
Raw datasets are stored in Amazon S3, transformed with dbt following the Medallion Architecture, queried through Amazon Athena, and visualized using Evidence.dev. The final dashboard is deployed as a static website on Vercel.
Understanding Airbnb pricing trends requires more than simply querying raw CSV files.
This project demonstrates how modern Data Engineering practices can be applied to build a scalable analytics platform by combining:
- Medallion Architecture
- Dimensional Modeling
- Data Quality with dbt
- Serverless Analytics with Athena
- BI-as-Code with Evidence.dev
The final result is an interactive dashboard capable of exploring prices, neighborhoods, and review scores with low latency and minimal cloud costs.
- End-to-end analytics pipeline
- Medallion Architecture (Bronze, Silver, Gold)
- Data modeling with dbt
- Data quality tests
- Amazon Athena integration
- DuckDB local caching
- Interactive dashboard with Evidence.dev
- Static deployment on Vercel
- BI-as-Code approach
The project follows the Medallion Architecture:
- Bronze: raw CSV/Parquet files stored in Amazon S3.
- Silver: cleaned and standardized datasets with proper data types.
- Gold: dimensional models optimized for business analytics.
dbt is responsible for:
- Cleaning price values
- Removing special characters
- Data type enforcement
- Integrity tests
- Dimensional modeling
Evidence.dev queries Amazon Athena only once to retrieve aggregated datasets.
The data is cached locally in DuckDB using:
npm run sourcesThis dramatically reduces AWS query costs while allowing millisecond dashboard performance.
Evidence.dev generates static pages that are deployed on Vercel while still supporting interactive components such as:
- BigValue
- Histograms
- Scatter Plots
- Dropdown Filters
| Technology | Purpose |
|---|---|
| Amazon S3 | Raw data storage |
| AWS Glue Data Catalog | Metadata catalog |
| Amazon Athena | Serverless SQL engine |
| dbt (dbt-athena-community) | Data transformation |
| SQL | Data modeling |
| DuckDB | Local analytical cache |
| Evidence.dev | BI-as-Code dashboard |
| Svelte | Frontend framework |
| Vercel | Static deployment |
airbnb/
│
├── models/
│ ├── staging/ → Silver Layer
│ └── marts/ → Gold Layer
│
├── analyses/ → Data Quality Checks
├── macros/ → Reusable SQL
│
├── reports/ → Evidence.dev Dashboard
│ ├── pages/
│ └── sources/
│
├── dbt_project.yml
├── package.yml
└── README.md
flowchart LR
A[Inside Airbnb Dataset<br/>Raw CSV Files]
B[Bronze Layer<br/>AWS S3<br/>Raw Data]
C[Silver Layer<br/>dbt + Athena<br/>Cleaning & Data Quality]
D[Gold Layer<br/>Dimensional Models<br/>Analytics Ready]
E[Evidence.dev<br/>Athena Queries + DuckDB Cache]
F[Interactive Dashboard<br/>Charts & KPIs]
G[Vercel Deployment<br/>Static Website]
A --> B
B --> C
C --> D
D --> E
E --> F
F --> G
- Python 3.9+
- Node.js 18+
- AWS CLI configured
- dbt-athena-community
aws configurepython3 -m venv .venv
source .venv/bin/activate
pip install dbt-athena-community
dbt deps
dbt runcd reports
npm install
npm run sources
npm run devOpen:
http://localhost:3000
npm run build
npm run previewThis project uses the São Paulo Airbnb dataset provided by Inside Airbnb.
- Bronze layer
- Silver transformations
- Gold dimensional models
- dbt data quality tests
- Athena integration
- DuckDB local cache
- Evidence.dev dashboard
- Vercel deployment
- Incremental dbt models
- CI/CD with GitHub Actions
- Dockerized local environment
- Scheduled dbt jobs
- Dashboard authentication
This project is licensed under the MIT License. See the LICENSE file for details.