Analyze and forecast demand in construction projects using Machine Learning

A construction project has between 8-10 deliveries per day on average, and transport accounts for about 10% of the emissions during a construction project. In order to reduce the environmental impact and increase transport efficiency by reducing the number of transports, the construction industry can work to coordinate transports to different construction projects. A problem today is that the construction industry consists of many small players who each order their material, which makes it difficult for the building material dealers to understand which deliveries can be coordinated. Furthermore, the data that is received via the orders is not always linked so that it is possible to see this coordination potential. Ahlsell wants to improve its transport fill rate, but in order to do this, it needs to link current data on deliveries to different construction projects to enable pooling. With the help of machine learning, order data will be analyzed and clustered to find different ways to identify patterns in the transports to construction projects which can be used to forecast the transport needs during different phases of a construction project. It will enable Ahlsell to be able to improve its fill rate and optimize its warehouses while serving the customer in a good way. This also makes it possible to reduce the environmental impact through fewer transports.

Goal

Identify transport patterns for different types of materials/projects in Östergötland. How to identify patterns in the data/clustering? How can one forecast future transportation and material needs for various projects based on the patterns?