Imagine that you are the first data scientist in a company, maybe in the industrial field, in one of the old industries or old economy branches. Then, you are a unicorn. Basically, you start data science from scratch: you must introduce, explain, promote, and establish data science. To manage this challenging task, dare to start simple!
Three years ago, I was exactly there – the first and for some time the only data scientist in the traditional, old industry company. The challenge was to start. They all say, start with the low hanging fruits. For a data scientist, the low hanging fruits are KPIs and other metrics, reports, dashboards, historical analysis. But here is the thing. On one hand you, with your ideas about machine learning projects, your training on advanced, fancy methods which you are really looking forward to applying. On the other, the reality of progress but not completed the digital transformation, messy data, people do not fully understand what your job is, do not know what is possible. To overcome this low-hanging-fruits dilemma, I needed to get rid of my, a bit snobbish, academic glasses.
There are a lot of reasons to start with low hanging fruits. Above all and the most reason is that these use cases help you to understand the business model. This is the very first thing each data scientist should always approach at any company. Taking a deep dive into the business model helps you to access where is your contribution, where are the high impact use cases and projects, and where there is business value. Sometimes applied method equals business value. But unfortunately for us, method affine people, the organizational needs of a company determine solely the business value of any project or use case, not by the method. Applications like dashboards and reports are relatively simple to produce and have a high impact on the company. Moreover, they are easy to communicate. Especially, if your company is at the beginning of the data science journey, the combination of high business value and good communication of the project will help you to promote and establish data science. People get to know you, you build a network, and you build trust with your stakeholders. This will bring data science forward at your company, together with your personal success and forthcoming sophisticated projects.
Editor’s note: Katharina is a speaker for ODSC Europe 2020. Check out her talk, “Dare to Start Simple,” there!
More on the speaker/author:
Dr. Katharina Glass is a data scientist and digital transformation enthusiast, employed as a manager data scientist at Aurubis AG, Europe’s biggest multimetal producer and #3 in the world. She is an empirical econometrician, was an empirical researcher for 10 years with a focus on forecasting and uncertainty. Since 2016, she matched both the science and industry worlds together. She has expertise in statistics, machine learning, and AI, among others. The portfolio of her projects is very broad – from KPI (and other reports) assessment, development and deployment projects, to deep learning models and AI development and implementation for relevant business cases.