We started by taking steps to identify the determining factors in causing algal blooms and making feasible initiatives towards predicting future occurrences. By using NASA APIs and datasets, we were able to train a machine learning model on the various factors that contribute to the formation of algae. We learned the importance of each factor.

We chose to place our focus on 120,000 km squared of water area in the Gulf of Maine, located off our East Coast. The main API used from NASA was the ocean chlorophyll concentration data from NASA’s Moderate Resolution Imaging Spectrometer. This was incorporated in the machine learning system that was given several hundred pieces of information.

With data for the entirety of our chosen area on a variety of factors including salinity, temperature, and chlorophyll concentration, we were able to train a machine learning model on it and determine the feature importance of the various factors. A link to our code can be found in the link below. The application prototype incorporates the crowd-sourcing techniques with a simple and intuitive user interface. Users can share their location and display the 4 conditions that influence algae bloom, chosen by our team, by pressing the “Ping” button.

One of the biggest challenges our group faced was prioritizing feasibility while designing our solution. We struggled finding something innovative and achievable. We overcame this by considering various things such as costs, availability and audience in the brainstorming process itself.

Link to code: https://github.com/Uzair5162/sustainPlus/blob/master/code.py