GreenGrid

Inspiration

Singapore's only landfill, Semakau Landfill, is projected to reach capacity by 2035, while our household recycling rate remains stuck at 11%. Singapore publishes rich open datasets on bins, e-waste points, and waste statistics, but they exist in silos. We were inspired to build a platform that connects these dots so planners don't have to guess where infrastructure is needed most.

What We Plan to Build

GreenGrid (ZeroWaste Navigator) is a Databricks-native intelligence platform designed to support circular economy planning:

1. Coverage Heatmap: Spatially joining open bin locations with URA boundaries and population data to identify underserved neighbourhoods (bins per 1,000 residents).
2. 2030 Trajectory Model: A time-series forecast using Facebook Prophet on 20+ years of waste history to quantify the gap to Singapore's Zero Waste Masterplan targets.
3. Intervention Simulator: A scenario tool allowing planners to test the impact of deploying new bins or campaigns before spending public funds.

Our Approach & Databricks Architecture

We plan to build this using Databricks' Medallion Architecture:

• Bronze: Ingesting raw open datasets from data.gov.sg into Delta Lake.
• Silver: Cleaning coordinates and running PySpark geospatial joins against URA planning boundaries.
• Gold: Computing opportunity rankings and forecasting outputs.
• Consumption: An interactive Streamlit dashboard on Databricks Apps, plus a Databricks AI/BI Genie Space for natural language policy queries.

What We Hope to Learn

We are excited to learn how to operationalize Databricks Unity Catalog, explore geospatial indexing (H3) on Apache Spark, and see how open data can directly inform public sector sustainability decisions.

Built With

• databricks
• delta-lake
• open-data
• prophet
• pyspark
• python
• streamlit