Posted on 12/09/2019 via BSH Home Appliances Group

What does a data scientist actually do?

We are pleased to introduce you to our colleague Anil Inal, the first data scientist in our supply chain organization. Anil uses algorithms to define sales forecasts with greater precision and formerly wrote his master’s thesis on this subject at BSH. But let him explain his profession to you himself:

Anil, what does a data scientist do? And what do you do with us in the supply chain department?
Generally speaking, what we do is try to find specific answers to very specific questions from what is perhaps initially a confusing wealth of data. For example, a self-driving car collects a vast amount of data during the journey. Algorithms can then be used to evaluate this data. This allows decisions to be made autonomously in different traffic situations, with the result that a self-driving car can constantly “think for itself”.

I am currently working on integrated business planning at BSH as part of an end-to-end global supply chain project called IMPulse specifically in the area of analytics. My primary task is to develop, expand and ultimately roll out the statistical sales planning process. I use algorithms every day in this context in exactly the same way.

And what exactly does that look like?
I split our products into different groups, for example, into appliances that sell extremely well and those that are more difficult to sell. We refer to this as segmenting. We look for the right algorithm for each group and this algorithm then analyzes the history of the product. More specifically, sales history is analyzed based on conspicuous patterns. The algorithm then uses these patterns to predict how the buyers in this product group will behave over the coming months. The sales planners are given this information and can adjust their planning as required.

Is that not still a bit up in the air, or are we already using this type of planning?
Our sales planners in Northern Europe have been using this method since the beginning of July 2018. There were further roll outs to other European countries this year and there will be new ones next year.
We refer to this at BSH as statistical forecasting. We worked together closely with key users in the development phase in order to also properly understand the needs of the user. We used agile methods, which helped us to also act in a really customer-centric manner internally.

As a data scientist, where do you see further potential for improvement within the supply chain?
I also worked with IT on an option to have our weekly sales reporting generated automatically. But there are possibilities elsewhere as well, not just in Sales. There are possibilities elsewhere as well, not just in Sales. I am certain that there is still enough potential throughout the entire supply chain for us to achieve further improvements with the aid of data science.

Thank you, Anil for the exciting insight into your tasks and the world of a data scientist.