A full-stack weather accessibility platform that generates narrated videos containing AI-narrated weather products and visualizations. The project demonstrates expertise in designing scalable APIs, implementing sophisticated backend processing workflows, integrating external data sources, and leveraging cloud services for production infrastructure.
Production Website: www.olievortex.com
Problem Solved: Provides weather information and severe weather alerts to visually impaired users, removing barriers to accessing critical safety information that is typically visual-only.
Core Capabilities:
- Narrated Storm Briefings - Automatically generates audio-narrated videos from Storm Prediction Center (SPC) raw text and imagery, delivering professional weather analysis in an accessible format
- Accessible Weather Maps - Produces weather visualizations optimized for colorblind users, ensuring weather patterns and hazards are distinguishable to users with color vision deficiency
- Severe Weather Archive - Maintains a searchable historical archive of significant weather events with comprehensive descriptions and satellite imagery, enabling users to research past conditions and trends
- Utilizes Google's text-to-speech API to narrate Storm Prediction Center storm outlooks.
- Backend process that combines C#, Python, and ffmpeg in one worksteram to generate videos and weather maps.
- Utilizes Azure Service Bus to queue time consuming processes that hydrate satellite imagery.
- Places a permenant copy of all ingested data into a Azure Blob container, following the medalian architecture (bronze-silver-gold).
- Create REST API to support the main website.
- Integrate websites and console applications with Application Insights to capure all telemetry.
- Greater than 80% code coverage
- Utilize Akamai Cloud to host MySql databases and Linode Compute instances.
For deployment and setup instructions: