Speeding up the time to insight for supply chains and logistics

Adrian Kosowski, Chief Product Officer at Pathway, was invited to the Data Engineering Podcast to discuss speeding up the time to insight for supply chains and logistics.

Some of our key takeaways from the conversation:

• Supply chain data analytics

In logistics and supply chain, the world used to be a more stable place where people planned two or three years ahead. These days it's one month ahead at best and things can sometimes change every week or even more frequently. A challenge then is differentiating between an anomaly and something which is different but that we would rather consider being “The New Normal”, e.g. a warehouse changing the location of its entry door. If you were to analyze the process using historical data, you would have a completely different view of waiting times and overall process sequencing. The same happens across all operations.

• Supply chain Management

Today, we see a lot of interest in ensuring that the industry can have human-in-the-loop interactions with existing systems. Pathway answers this market need by enabling input thanks to its reactive design. The user can tune input parameters, make corrections to input data points, and add some settings, guidelines, hints, suggestions, and improvements. They can add them into the system through a user interface and then see how those fixes are being taken into account in realtime.

• A discrete world

One of the leitmotifs behind Pathway and I’m disclosing here some of our secret sauce is that we like discrete worlds and discrete data structures. One of the reasons is that when you are updating things in a computer, the cost to update a memory cell or a data entry is the same regardless of whether you are updating the value by one or by a lot. In some sense working in a discrete world means that you want to work with discrete changes. Working in a continuous world would be like saying “oh there was just this little bit of change”. We like to work with discrete data which means we put data as soon as possible in a discrete representation.

Watch the podcast recording here:
https://www.dataengineeringpodcast.com/pathway-database-that-thinks-episode-334/