Maritime
Machine Learning - Forecasting Singapore's Container Output
Model Accuracy: 96%
Tags: #Maritime #Regression #Forecasting #Singapore #DataVisualisation
Background
Singaporean port handle thousands of large container daily connecting between large ship carriers and the port.
The average cargo ship can carry around 18,000 containers (source) which means that maritime transportation is a cost-efficient way to transport high quantities of containers over large distances. However, due to the large quantities and size of goods delivered will cost time and productivity. Forecasting is one of the method to minimize those costs. Forecasting gets the business to look for a habit at the past and real-time data to predict future demand (source). Moreover, doing a forecast is time-consuming and resource-intensive as the data from port always grows exponentially in daily basis.
Solution
By setting up Machine Learning models to make a forecast, Analytico Asia optimized a prediction based on historical data and real-time data. While the prediction is never 100%, the accuracy of our model is so accurate that only 4.6% of the forecast is false.
Analytico Asia provides some valuable insight for the port business, opportunities to course-correct past mistakes and adjust future approaches. By doing so, the port can saves cost and time for doing the forecasting while increasing productivity respectively.