Predictive Solutions that Make the World Better

Geospatial Aedes Hotspots

The WHO estimates that 2 out of 5 people worldwide are at risk of infection with dengue. Dengue is an incurable disease which could lead to death, especially for children and the elderly. The Aedes aegypti mosquito is the primary vector of dengue, occurs mostly at localized urban tropical areas.

Challenge

What if AI could help organizations eradicate dengue effectively before it happens? How can AI produce early warning to make preventive actions?

Our Solution

The Terra-AI.SG team has come up with first-in-the-world AI-based aedes hotspot prediction which can identify potential dengue hotspots up to 12 weeks ahead of time. This webapp called Eradicate Dengue enables targeted preventive actions which significantly reduces the cost of such operations while also lower dengue occurrences.

This solution is powered by Autocaffe, which allows our team to run thousands of experiments with different model configurations automatically.


Vessel Estimated Time Arrival (ETA) Prediction

90% of the world's trade is conducted by sea transportation. Yet, last year, more than 25 percent of vessel arrivals were delayed more than 12 hours. Various factors such as weather, berth availability, port congestion and delays in downstream ports play important part in determining the arrival time of the vessels. Delay in vessel arrival results in billions of wasted dollars annually due to disrupted scheduling and port congestion.

Challenges

What if AI could predict when exactly the ships would arrive at any ports? How could AI integrate and process all information to come up with accurate prediction?

Our Solution

Terra-AI.SG is developing an interactive vessel ETA calculator, which predicts the arrival time and congestion levels of ports. We are currently launching this solution for several major ports in Indonesia.

Our solution is powered by the Autocaffe learning framework and the Smojo Programming Language. It runs data pre-processing and classification seamlessly over hundred millions of data points, before being sent to the machine learning framework for accurate prediction and analysis.