Oxalgae Data Solutions - Space Apps Challenge

Oxalgae Data Solutions| To Bloom or Not to Bloom

Project Details

The Challenge | To Bloom or Not to Bloom

Your challenge is to solve the mystery behind algal blooms! What factors cause blooms in some water bodies but not others, and how can we better predict their occurrence to prevent harm to aquatic and human life?

Determination of factors influencing the growth of algae in lakes, using Machine Learning algorithms

In this project we decided to take the task of determination of factors, influencing the occurrence of algae blooms in lakes and rivers. We will evaluate to which degree different factors influence the occurrence of blooms and how we can deal with them.

Background

I’ve been always passionate about the idea of algae blooms and wondered what are the reasons behind the occurrence of this natural phenomena. In order to find the answer to this question, I moved further and decided to use my data science and machine learning skills to find them.

What it does?

We decided to find connections and hidden patterns in data on concentration of chlorophyll (which equals to concentration of algae in our case) and geophysical data in order to find, how different parameters influence the growth of algae. However, it would be a teadous task to analyze the whole data for each lake and sea on Earth, therefore we decided to take Ontario and Erie lakes as test-subjects for our research. Unlucky we haven’t had much time to prepare and develop our solution, but we managed to use approximate analysis on chlorophyll concentration, available for photosynthesis radiation, biologically-available carbon-sources and other parameters. After research on data (primary one), we came to conclusion that available for photosynthesis radiation, landscape data, biologically available carbon sources and temperature influence the growth of algae the most. The degree of influence is the next:

Biologically available cargo sources
Available for photosynthesis radiation
Temperature
Landscape

NASA resources

In our research we used the next NASA resources:

We used data on chlorophyll concentration and geophysical parameters provided by MODIS (moderate resolution image spectroradiometer) provided by Ocean Color service - modis.gsfc.nasa.gov, https://oceancolor.gsfc.nasa.gov/
We used Panoply, developed by Goddard Institute for Space Studies, a powerful tool for quick exploring the data, stored in netCDF format, and generating plots - https://www.giss.nasa.gov/tools/panoply/
We also used data on chlorophyll concentration from MODIS provided by NASA NEO - https://neo.sci.gsfc.nasa.gov/view.php?datasetId=M...

Space Apps Offers

Sadly we haven't used any SpaceApps offers in our project, but we are going to explore and use them next year.

Future plans

We are definitely going to continue work on our project, namely we are going to apply different machine learning techniques (Polynomial regression, clusterfication, et cetera) in order to evaluate the numerical and exact value for influence and build a tool able to predict the growth of algae and the chance of occurence of algae bloom in specific area.

Built with

We used Panoply (written on Java) as the main tool for exploring the data together with data provided by Ocean Color from MODIC satellites. Graphics were made in Paint.NET, free graphics designer and Figma. In the future development we are going to use Python as the main data analysis tool.

Try it out

You can try to detect those patterns by yourself by looking at graphs which we used in our research. The album with them can be found here: https://imgur.com/a/EwL6tZ2