Blooming Cosmo| To Bloom or Not to Bloom
Project Details
The Challenge | To Bloom or Not to Bloom
Predict Algae Bloom with Machine Learning
We wish to train an M.L. algorithm with remote-sensing data from NASA satellites to output the probability vector for the next week for a bloom to occur.
Summary:
We have developed a document that explains the background of algae blooms and possible sources and devised a prevention idea. Because we do not know for sure the cause for these blooms we decided to let an machine learning algorithm to find patterns and chose what information is relevant. We feed it with as much satellite data (surface temperature, chlorophyll mass / cubic centimeter, satellite images etc.) and he will decide after it's training on what data to look and which to ignore.
This is the documentwith a more detailed description: tiny.cc/sgguez
One of the challenges for this approach would be to complete the data gaps that appear from satellite readings because of cloud coverage at the pass of the satellite and also reading errors from the sensors.
Resources used:
In our research and documentation we have used the internet vastly and the provided links in the challange page were extremely helpful. A few links would be:
More links for resources used can be found in the document provided (tiny.cc/sgguez) in the bibliography section.