This is the third and final article on Data Analytics and AI in Transport & Logistics.
- Journey
In this sector, one of the biggest challenges regards the transport of goods itself. There is a constant concern to keep the flow of journeys running continuously and smoothly. Choosing the best routes, selecting locations with suitable parking for long vehicles or bypassing building works and chaotic traffic situations are examples of challenges faced by a carrier. Here’s how Data Analytics and Artificial Intelligence can help.
Route Optimization
Technique that allows determining the best route to a particular destination. The “best” route is not always the shortest: we may prefer it to be the fastest, the most economical or the most comfortable, for example.
It is paramount in the transport sector whenever different vehicles need to travel between goods delivery points. A non-optimized route can represent high costs in terms of time, fuel or vehicle wear.
When optimizing routes, in addition to the distance between stopping points, attention must be paid to vehicle characteristics, loading and unloading times, traffic, drivers’ rest, among many other factors. Through Machine Learning, it is possible to analyze all these variables in an integrated and longitudinal way (over time) in order to predict and recommend the best travel route.
Web Scraping
Through Web Scraping solutions, it is possible to access websites in order to extract information that is then stored in a structured or semi-structured way.
Web Scraping’s usefulness can be seen, for example, in capturing contacts to generate leads, in collecting prices of competitive products for comparison and monitoring, or simply in capturing published reviews about a restaurant on websites such as Tripadvisor.
Usually, data from multiple sources is stored in a single location (e.g., Data Warehouse / Data Lake), which facilitates data scientists’ activity in data analysis. Although it seems to be a “distant” technique from transport and logistics, the methodology allows feeding machine learning systems that then offer solutions to difficulties felt by fleet and route manager. Through Web Scraping it is possible to collect up-to-the-minute information from the web about the weather, accidents on the road or even fuel prices at gas stations, making suggestions for optimizing routes, especially in long-distance transport.
- Destination
In transport and logistics, a destination is often just the beginning of a new journey, especially in multimodal transport, where goods can be transported using different means of transport (airplane, ship, train, truck…). Along these interim destinations, there is a need for transshipment, loading/unloading operations, temporary storage, verification of the integrity of goods, etc. Here are some solutions:
Computer Vision
This component of Artificial Intelligence allows computers to “understand” and recognize visual information (for example objects or people’s images or videos) after being trained to do so. With the inputs received, they can then suggest actions or flag anomalies that can be corrected.
It is a rapidly growing component especially due to its versatility – it can be applied in multiple situations. Imagine a logistics center: through Computer Vision systems it is possible for trucks from different sources to be “recognized” at the entrance of warehouses without human verification, automatically knowing the truck’s identification and, through that, its load, warehouse it should head to, as well as its assigned driver.
Another example would be its inclusion in conveyor belts, where, through Computer Vision there would be an automatic detection of defects/deteriorated goods. In this way, it would be possible to give prior notice to those in charge and, therefore, to make corrective decisions.
Finally, Computer Vision can also help in fraud detection, as it allows identifying people’s access to goods (i.e., who had access and to which goods).
Bag of Words
Bag of words is a natural language processing (NLP) technique that detects the existence and frequency of keywords in texts or documents. The presence of these keywords can facilitate the classification of these documents automatically, with no need for human intervention.
This can be done at the visual level – instead of words extracted from text, there are image excerpts taken from larger images. Whether text or image, elements are always converted into numbers, and can be analyzed in a computational context.
Its use is important for classifying documents, images, detecting objects (e.g., detecting cars in an image), determining distances, capturing license plates through radars or even recognizing events (e.g., road accidents).
In transport and logistics, it has a wide range applications. Due to the presence and frequency of certain keywords in a document, it is possible to automatically classify it as belonging to a specific category (e.g., a document is categorized as a waybill).
Let’s look at another hypothetical example: a robotic storage system has to stack, among a set of boxes, only those with a specific icon on the packaging. Through the Bag of Words technique it is possible to detect the icon, even if there are many other indications in it. The robot will only “select” the one with the intended icon.
To sum up, Data Science and Artificial Intelligence have a lot to offer to transport and logistics. It should be noted that the solutions presented are just a tiny part of all existing techniques. The important thing is that decision-makers establish contact with specialists in order to achieve tailor-made solutions. It is known that companies are not all the same and have different pain points. Data Analytics and Artificial Intelligence offer many solutions with proven effectiveness.
Inspiring companies to explore new challenges through the power of data