Professional ML Team

Universal Event | Leveraging AI/ML for plastic marine debris

NASA Project; Plastic Marine Debris Classification-Machine Learning Software

High-Level Project Summary

With this project, plastic, metal, etc. found in marine debris. I have provided the classification of derivative wastes according to various parameters (date, country, etc.). The machine learning software I have created works with high accuracy (The highest classification model accuracy rate is about 97%). The project regularizes the irregularity of the various data and solves the complex plastic proportions in which country they occur, in what year.

Link to Project "Demo"

https://github.com/emirhanai/NASA-Project-Plastic-Marine-Debris-Classification-Machine-Learning-Software/blob/main/NASA%20Project%3B%20Plastic%20Marine%20Debris%20Classification-Machine%20Learning%20Software.pdf

Link to Final Project

https://github.com/emirhanai/NASA-Project-Plastic-Marine-Debris-Classification-Machine-Learning-Software

Detailed Project Description

NASA Project; I Developed Plastic Marine Wreck Classification-Machine Learning Software. Plastic, metal, etc. in this software. various waste and garbage; Classified by seasons, by photos they have, by country, by date (year) and shoreline name, with high accuracy, precision, sharpness and f1 results. The models were carefully prepared and examined one by one. Damages in the data have been corrected and made suitable for artificial intelligence. Some (.dot) files have a high number of megabytes and strings, so (.png) was uploaded unformatted due to my insufficient resources. This software has been prepared by me personally for the NASA Project.

Space Agency Data

I have received data from the Ocean Conservancy. I used them in .CSV format and completed the integration process with the PANDAS library in Python.


These data have given my project a very comprehensive direction. Multiple date types and start/end times are also given. Although the data is very detailed, it is also irregular. I made the data suitable for artificial intelligence by shaping it manually (without breaking its structure).

Hackathon Journey

I took part in all the processes of the project as the only member of my team. I tried to organize the data like a data scientist and clung more to the numpy-pandas libraries. In this way, I learned in detail the feature of using the to_string function in the pandas library. The biggest feature in my choice of this challenge was that it was a preliminary step towards reducing various marine pollutions found in the world. I have started this preliminary step. I tried to solve all the problems in my team by myself. I neuronally mapped out the business process and flowchart in my head and automated it. In addition, I have successfully concluded by thinking and researching software errors.

One person I want to thank is 'ELON MUSK'. Thanks to his advice, I am trying to become an expert in machine learning and I am learning a lot of knowledge and experience by taking advantage of the fact that the information is free!

References

Data Source: https://cscloud-ec2020.opendata.arcgis.com/datasets/data-marine-debris-monitoring-and-assessment-project-mdmap-accumulation-report-plastic-pollution/explore


Tools:

  1. Python
  2. NumPy Library
  3. Pandas Library
  4. Scikit-Learn (Sklearn) Library
  5. And various methods in Library

Tags

machine learning, artificial intelligence, ml, marine debris, classification

Leveraging AI/ML for plastic marine debris

Marine debris is one of the most pervasive threats to the health of coastal areas, oceans, and waterways. Your challenge is to leverage Artificial Intelligence/Machine Learning to monitor, detect, and quantify plastic pollution and increase our understanding about using these techniques for this purpose.