Coral Bleaching Prediction

Inspiration:

Our inspiration comes from a famous news article where more than 40% of corals around the world famous Great Barrier Reef, home to one-third of the world's population of corals were reportedly bleached. This incident scared us as corals reefs are vital ecosystems that support marine biodiversity, protecting coastlines floods, and sustains millions of people worldwide. The potential loss of these vibrant underwater habitats motivated our group to take action and create an innovative solution to predict and prevent coral bleaching before its too late. Through this, we can save these important underwater habitats for future generations to come.

What it does:

Our solution, which is a model that predicts coral bleaching, takes in multiple input parameters like depth of coral and proximity to human activity (e.g., the closest town). Using these input parameters, the model outputs the percent bleaching of the coral and rates it from low for percentages from 0-25, medium from 26-60 and high from 60-100. This categorization helps to give immediate attention needed for coral areas and can help to allow for more efficient conversation efforts for coral reefs in different areas. This will protect vulnerable coral reefs before they are irreversibly damaged from bleaching.

How we built it:

We built it using Jupyter notebook where modules using python language like numpy, pandas, matplotlib, seaborn, scikit learn are avaliable there. We used numpy and pandas to manipulate and clean data, matplotlib and seaborn to create meaningful visualisation of data and scikit learn to make models for predictions.

After using Jupyter notebook, we used streamlit to host a website containing our predictive model and also offered a tutorial on how to use the website model to predict bleaching percentage from taking the best model from our model training in Jupyter notebook.

Challenges we ran into:

We ran into some issues as we forgot some fundamentals of machine learning, which lead to us creating inaccurate models.

We also ran into designing issues on streamlit as CSS isnt really our strong suits for our group.

Through hard work and determination, we managed to overcome this issues and come up with our solution.

Accomplishments that we're proud of:

We are proud of creating our solution as this is one of the first few times we were given freedom to choose a topic. This opportunity gave us a chance to try something new and accomplish something new and innovative.

We are also proud of our group worked together as we rarely get chances to work outside of school together. This project strengthened our teamwork skills and allowed us to combine our unique strengths to achieve a common goal of making this innovative solution.

What we learned:

We learned how to do project management in codes, using GitHub to collaborate with one another at the same time, which made our progress in this solution more efficient and readable.

What's next for Coral Bleaching Prediction

To further enhance our solution, we plan to implement the usage of IoT sensors to predict the coral bleaching percentages of coral in different regions like the Indian or Pacific ocean, enabling real-time data collection on critical environmental factors for corals reefs like water temperature and acidity.

By leveraging the usage of IoT sensors, it allows for our solution to be more efficient in finding corals with high bleaching percentages across multiple locations and allow for immediate and targeted conservation action to be taken to protect the coral reefs.

The usage of IoT sensors will help us to create a scalable, real-time monitoring system that empowers efficient conservation of vulnerable coral reefs around South East Asia.