ORBITAIR — AI-Powered AQI Forecasting
NASA Space Apps Challenge 2025 — Top 5 in India
01_MISSION_OBJECTIVE_AND_CONTEXT
> Mission Objective
Developed a FastAPI backend with TimescaleDB for geospatial AQI forecasting. Integrated NASA TEMPO, EPA/OpenAQ, and NOAA data feeds. Built a forecasting pipeline achieving 98% prediction accuracy. Created a React and Leaflet dashboard for pollution visualization and explainable AI outputs.
> Problem Statement
Aggregating geospatial data from NASA TEMPO satellites, NOAA feeds, and EPA sensors in real-time, and forecasting AQI with spatial accuracy, is computationally resource-intensive.
02_ENGINEERED_SYSTEM_AND_ARCHITECTURE
> Implemented Solution
Built a FastAPI time-series backend backed by TimescaleDB to index high-volume geographical points. Created a forecasting pipeline that parses sensor datasets and trains models to predict AQI, rendering outcomes on a Leaflet map.
> Architectural Design Schematic
FastAPI -> TimescaleDB Time-Series Indexes -> AQI ML Forecasting Engine -> React + Leaflet UI.
03_OPERATIONAL_CHALLENGES_LOGGED
- Handling dynamic time-series data ingest pipelines without causing write-locks on the databases.
- Mapping complex multi-dimensional geospatial metrics onto a responsive 2D Leaflet canvas.
04_METRICS_AND_DEBRIEFING_ANALYTICS
> Mission Results
- Ranked Top 5 in India in the NASA Space Apps Challenge 2025 (competing against 823 teams).
- Achieved a 98% time-series prediction accuracy rate on forecast AQI index points.
> Engineering Lessons Learned
- TimescaleDB hypertable partitioning is highly effective for running geospatial queries on large datasets.
- Explainable AI outputs are critical for helping city planners trust predictive environmental models.
05_LINKED_GITHUB_REPOSITORY_DOSSIER
- Time-Series Geospatial Partitioning
- Polyglot Tech Stack Ingestion
- Geospatial Hypertable Scaling
> Live README.md Documentation
ORBITAIR — AI-Powered AQI Forecasting
A geospatial forecasting platform that indexes satellite and local sensor data to predict air quality.
## Features
* Geospatial Ingestion: Integrates NASA TEMPO satellite and EPA/OpenAQ sensor feeds.
* High-Volume Time-Series: Backed by TimescaleDB hypertables.
* Explainable AI Dashboard: Beautiful React map rendering pollution forecasts.
MISSION_METADATA
TECHNOLOGY_INVENTORY
Secure integration path for ORACLE system analytics. Connects real-time compiler telemetry and AI metrics evaluation on this build.
Milestone tracking datastream for operational deployments. Maps historical project milestones, commit frequency, and release paths.
Candidate alignment report panel. Matches project challenges and achievements against targeted job skills matrices.