Report by Johan Hellman.
During my years as logistics manager at a large manufacturing industrial company, we often took a look at the automotive industry and the advanced logistics concepts developed and applied there. We also recurrently recruited logistics experts from the industry, in the hope that they could help us to develop and improve our own logistics. However, the problem that often arose was that the experts from the automotive industry were experts in logistics in the automotive industry. Accordingly, they were good at working in an advanced logistics system that was already functioning very smoothly. By contrast, they were not good at working with a more rudimentary system and developing it. Simply put, they were good at improving the running pace of an accomplished runner, but did not know how to teach someone to walk.
Visual data analytics
When it comes to tools for visual data analytics – or “business intelligence” as it is often called – three types of analysis are usually discussed: descriptive, predictive and prescriptive. Descriptive analysis concerns what has happened; predictive concerns what will happen; and prescriptive concerns proposing the optimal choices given the prevailing conditions. Of course, the prescriptive analysis is the main aim, and in this era of artificial intelligence and big data, the potential seems almost limitless. Advanced examples already exist in the transport industry in which the most well-known are perhaps Amazon’s patented concepts for shipping goods in advance based on expected orders (“anticipatory shipping”).
At Unifaun, we are currently well underway with the development of the next version of our offer for visual data analytics. It is clear that both we and our customers often face similar challenges to the ones described in the introduction. Tools and approaches are often adapted to further improve logistics systems that are already working smoothly – think, for example, about Amazon and their anticipatory “shipments”. However, many do not have an especially well-functioning system and exist in a much more disorganised and imperfect reality. Shipments are booked via many different channels and with many different carriers. It is far from clear that the carriers with whom agreements are signed are the ones that are used. Many also have difficulty in setting target dates for collection and delivery, or for that matter in being able to measure the actual outcomes. It goes without saying that it will be difficult to measure delivery precision (i.e. delivery punctuality) without knowing when the goods should have arrived or when they actually did arrive.
Overview and understanding
The conclusion we have drawn is that perhaps most important of all is to develop visual analysis tools that allow you to monitor and control the basic work of structuring and organising your shipments. For example, it can be a question of visualisation that makes it easy to understand the nodes between which there are shipments, if any at all, in the transport management system used, and with which carriers and within which weight ranges. A second example could be visualisations that make it easy to understand the appearance of the status flow and status quality from different carriers. A third example is the facility to study shipments, weights and transport costs related to different carriers, services and country relationships.
Finding connections in large, complex data sets
The strength of modern tools for visual data analytics is that they can so easily and quickly identify interesting relationships in complex data sets. Clearly showing the situation as a whole, while being able to dig deep into the details and to cut the data along many different dimensions, creates good conditions for driving forward improvements by using the right measures in the right areas. Our mantra is to help the customer get an overview and control to be able to work with continuous improvement. We consider visual analysis to be a prerequisite for success in this. Perhaps it is even the case that the great value in visual analysis lies in the journey itself to the exact and effective reality that is so often seen as the starting point for more advanced applications for predictive and prescriptive analytics.