The app is easy to use with a flexible data model. Users can log in and have a record of all their past observations. They can create new ones, or browse the previous ones.

Another value that this app brings to the users is effective notifications. When the user creates an observation entity, they'll be able to define a communication channel, lead time, and a list of recipients. Once the datasets are processed and loaded by the providers, they'll be notified again by the system via email. They can click on a button and go directly to the respective observation details.

The app uses a queue worker for these transactions to maintain its performance.

The system also has a Python API (made by FastAPI) consumed by the web-based app. The API contains endpoints for calculating overpass time (Using SkyField and datasets by CeleStrak), drawing maps (Folium, Geehydro), and getting metadata for a specific scene/pixel by GEE's Python SDK.

The app also has two admin commands "app:send-email" and "app:check-data" for event notifications (e.g., SatelliteApproaching and DataReady) and data-availability checks respectively. These two commands are scheduled to run daily.

I've also built the code to add NDVI bands to Landsat images, but I excluded the code since it wasn't part of the scope.

Another use case would be creating specialized workflows for different research teams working on various projects on agriculture, oceans, or jungles.

Software Stack

I used the TALL (Tailwind CSS, Alpine, Laravel, Livewire) stack for this project with a bit of FilamentPHP since it's a great tool for rapid prototyping. I also used Python's FastAPI to create a lightweight API for interacting with Google Earth Engine.

My Database was SQLite, but the app is database-agnostic by design.

To do:

-Dockerizing the project for easier dev environment setup

-Add Landsat 9/Sentinel support for displaying bands

-Adding multi-tenancy to the app so teams with multiple users can work on the same observation projects

-Adding Slack as a communication channel

-Add Latitude/longitude to the place search

-Add a cloud cover filter to the analysis panel

-Add the missing histogram

- Use USGS data to predict overpass time, and use the implemented algorithm as a fallback

- Add multiple dataset catalogs other than GEE.

- Being able to animate time-series data

Software Stack

I used the TALL (Tailwind CSS, Alpine, Laravel, Livewire) stack for this project, since it's a great choice for rapid prototyping.

2024 NASA Space Apps Challenge

Cosmic Radar