The problem Oil Spill Prediction and Prevention Dashboard solves

Problem

During the transportation of oils in large ships, there are large chances of oil leakage and spills. This causes serious effects on the environment and affects the water eco-system for the long term. Sometimes oil leakage can cause fire accidents as well.

• No mechanism to monitor the current state of the environment.
• There is no automatic alert system that can alert before an accident happens
• Lack of automated AI/ML mechanism that can predict the chances of oil spills.

Solution

Prevention is better than cure

We have developed an ML-based system that can use the current state of ships and other data like weather sea conditions to predict the chances of oil spills or leakage. The system can show any abnormal state in the system.

If there are chances for oil leakage the system can alert via its interface and shows the reason for which the damage occurs.

• An automated AI-based system that can detect abnormal behavior and give alerts and information of the current state.
• The dashboard shows the detailed analysis and the real-time prediction about the ship in a visual manner (graphs, charts, etc).
• The machine learning model can use a weather monitoring system in the ship or weather API for weather analysis.

More Features

• Collision Alert
• Live data of the surrounding ship available for every ship. Triggers alert system, if two ships come close to each other.
  Danger Prone Area Alert
• The system will alert the admin if the ship is heading towards a danger-prone area.

Challenges we ran into

When we decided on our idea we ran into one of the biggest problems immediately that's the data. For our work data is the key factor we took some time and searched a lot and we did get some data from internet.

But it's not over some data we need are from live information so we thought about how it's feasible to implement the solution while we need live data. So we did research (a lot) and read about the actual environment where the solution is implemented and we know what to do. We mocked the environment live feed of data and did it.