Machine Learning in Transportation

Machine Learning in Transportation 

Logistics and transportation are continuing to evolve as new technology enables drivers to have access to key information, companies to track shipments across the globe and accurate predictions to lead to a better margin. Machine Learning has the ability to quickly provide value in these areas and bring your transportation processes into something efficient. What processes do you use to predict the local demand for pickups?  If the process is manual or depends purely on expertise, you are creating risk for your business. Pariveda has developed a Machine Learning approach to logistics demand forecasting that enables operational efficiency. Read how we helped a client below and reach out to learn more. 

Predict Demand for Long Haul Moving to Price Dynamically

During peak moving seasons, the Atlas agent network of Atlas Van Lines will work together across markets to meet customer demand.  However, their ability to forecast capacity was manual and labor intensive, relying on the wisdom of people with many years of experience and their gut instinct. Atlas had the historical data from 2011 forward and desired to find a way to dynamically adjust capacity and price based on future market demands.

Using the AgileML Methodology ,Pariveda went through an identify and assess phase to ensure that the problem statement was clearly identified. They then assisted Atlas in building a Machine Learning model by implementing sophisticated feature engineering and training 252 individual models, which predict the demand (in lbs.) by region every day up to 6 weeks (42 days) into the future. Each day, the results of the predictions are displayed in a heatmap that enables quick line-of-sight into days with excess demand and days with excess capacity. The model is trained using the AWS SageMaker platform’s Jupyter Notebook Servers. Predictions are generated by loading and executing the 252 models in parallel on the AWS Lambda Step Functions Platform.  

The model allows the operations team to quickly see which days will require extra attention, resulting in a significant cost savings due to decreased shipment delays. Additionally, the Marketing team can see which days may require additional effort to secure enough orders to fill capacity.       

Pariveda's ML Methodology 

Our overarching principle in our AI/ML engagements is to help our clients build a sustainable AI/ML capability through a holistic view of the people, process, and technology required to support that capability. We partner with Amazon Web Services to leverage the latest technology and provide value. 


Let us show you how we can apply ML to your business needs in transportation. Email us at