Pet project to learn ML
Predict sunset by weather data
Input:
- Lat, Lng
- Weather data in place
- Weather data from 160km on sunset direction
Output:
- Is sunset is colored (red)
- Is the horizon clean (without clouds)
- Is horizon colored
- AWS
- ECS (Fargate) (To save streams > 30min)
- Lambda (Serverless server)
- EventBridge (To schedule recorder)
- ECR (For docker images)
- DynamoDB (To collect dataset)
- Terraform
- FFmpeg to download stream
- NestJS
- Build NestJS app
- Build Docker image
- Apply terraform configuration
- Collect data set from several places
- Create ML model to predict factors of a sunset