Machine Learning @Trukky

Trukky was introduced into the Indian marketing ( market) in January ( Early) 2014( 2015) with an(a) object (vision) to change and innovate ( disrupt & simplify) in the freight industry with our cutting edge technology and unparalleled customer experience. The road freight volume in India is forecast to be 2,211.24 billion freight ton per kilometer, growing at 4.7 percent. The road freight industry in India is known for its highly fragmented nature and is notorious for highly dynamic and opaque pricing. The spot market changes on a daily basis if not hourly and the prices are sensitive to multiple events like availability of vehicles, seasonality, industry, services, and other operational parameters.

Before we move ahead and explain the problem in the freight industry, our  Trukky’s USP is simply to provide an instant pricing service ( make price discovery hasslefree for shippers) and provide a single point digital solution to ship goods seamlessly without wasting billions of Man hours otherwise spent on to find the cost of shipping and then finding the right carrier for fulfillment to shippers and carriers before they book. Therefore, in a market place that is constantly evolving and changing, We at Trukky took it upon ourselves to address the price discovery challenge for shippers to increase their productivity decide how to make reduce the operational costs, and eradicate displace the highly fragmented and unorganized middle-man to help bring the fair price, convenience & improved services for our shipper’s customers. We have collated  huge amounts of metadata across multiple variables, and with curated data spread and machine learning processes our ML team makes sure that we take account of all of the dynamic variables, trukky not only became the first digital platform with price discovery but also one of the first digital logistics platform with data-driven decision making right from customer acquisition to fulfillment

In order to understand this article well, it would be good to introduce ourselves to the basic terminology in the logistics marketplace sector. Some such important terms are:

  • Demand – Customer order for material movement from the origin (O) to destination (D)
  • Supply – Availability of trucks with truck suppliers
  • Vehicle type – Classifications are typically made with regard to body type, length and carrying capacity of a truck
  • ODVT – A combination of origin source (O), destination (D) and vehicle type (VT)


Pricing and truck service available today is a major operational bottleneck and there are key reasons for the lack of organized stakeholders in this industry. Both tech and non-tech players find it difficult to operate and scale-up in the spot market, enabling the brokers to enjoy the majority of market share.

Trukky, an online transport service which predominantly operates in the spot market with instant pricing and truck cargo space availability across the trucking network, enables SME consumers to not only get competitive prices but also book transport service instantly without haggling with multiple brokers.

In addition to full truckload, Trukky offers customers a unique load-sharing service enabling consumers to transport shipments right from 100 Kg and pay only for space utilized in the consignment.


The price index named Trukky Rate Utility Index (TRaUI Index) is already live on the portal for consumers and industry experts to track & study daily changes in Freight rates and going forward the index would trend hourly rate changes.

Trukky Index is a unique initiative by Trukky which will not only allow a detailed analysis of fluctuation of Freight rates nationally but would also enable tech companies like Trukky to Optimize the Freight space. TRUI Index has been compiled with Data collated over 40 months covering 10000+ trips and has been fine-tuned to accommodate variance due to seasonality, freight type, type of vehicle, and supply/demand dynamics.

Trukky Rate Index works in real-time by analyzing supply, demand, and reference price feed from truckers on daily basis, and we are able to do it through connected transportation and bringing a basic system of demand and supply on a technological platform. 

Data Collection

User activities are critical data points for us. These include things like how much time they spend on monitoring the price on the platform, how many times they quote on a particular ODVT etc. These user activities are captured via as event handler system which then stories such events and data points in our database clusters.

Our data process includes filtering, organizing, and analyzing millions of data points using libraries like NumPy and Pandas in Python.

Data Modelling

From the beginning, the task at hand was as big as it seems, however, our most important step has been to clearly frame each prediction problem. This has allowed us to minimize bias and harmful feedback which might be detrimental to our original ML models. In some cases it is as easy as filtering some observations from training data, however, these decisions amplify our business impact exponentially.

We have experimented with various models ranging from gradient-based decision trees to deep neural networks. But for a given set of features, performance varied minimally across different models. We, therefore, focused on feature engineer to get the maximum performance of our models. 

Training and deploying so many models to serve different objectives across our platform is always too big of a challenge. However, when you have tools like Tensorflow, Hadoop, and Spark, you can focus on making your pipelines efficient and clean as possible. Moreover, we use Tableau for our internal data visualizations and evaluations.


Machine Learning at Trukky is not just about engineering a holistic pricing engine, we identified a plethora of problems in the logistics sector that can be solved using our techniques. Use cases such as dispatching drivers efficiently, customer matching, accurate ETA, matching demand and supply through incentive and promotions, fraud detections, and customer support were identified and have been our focus so that we improve our services as countless variables and constant change in the industry. 

Leave a Reply

Your email address will not be published. Required fields are marked *

You May Also Like