Data Science is maturing

Every company can work with Data Science, says Van Leeuwen, small or large. But she warns companies not to step into this field blindly to avoid failing projects. In 2017, Gartner stated that 85% of big data projects fail, a McKinsey survey from 2018 found that only 8% of 1,000 respondents with analytics initiatives engaged in effective scaling practices. In 2019 the tech news site VentureBoat asked the question ‘Why do 87% of data science projects never make it to production? "The reasons these projects fail are that the people doing the project are not trained well enough, the projects are not being internalised in the culture, and because of a lack of business adoption. Business adoption is an important aspect of data science to get running," emphasises Van Leeuwen.

One way to get data science from the ground is to start with the so-called low hanging fruit. Van Leeuwen gives an example. "We apply process mining in a lot in companies to get insights into what type of activities take the longest time in an organisation. This way you can easily pinpoint what is delaying the business process, where the most effort and time is going into, and identify bottlenecks. Then you can take direct actions to increase the business process efficiency. This is easier in organisations where digitalisation is already in place."

To Van Leeuwen the business value of data science is obvious. "The return on investment for data science is starting to pay off. Data Science can help retailers getting insights into consumer behaviour and purchases, it can help purchase managers forecasting the amount of stock that must be in storage, and it can help marketing managers to standardise pricing. Thus, data science can save money."

Besides goal, and business adoption, it is the people who determine the success of Data Science projects, says Van Leeuwen. These are off course the data scientists who understand, read and can work with data. "You also need business specialists who can translate the data insights to the business field. These two kinds of people live on two islands. Data scientists can make business better, but if business people do not understand the solution, just see the costs of it, you miss the business adoption. There are huge steps to be taken for business people to get more understanding of the capabilities of Data Science and why and how it is going to work for their business."

Based on the needs for understanding and applying data science, Van Leeuwen developed, in collaboration with the EIT Digital Professional School, the Applied Data Science course. The training is a combination of an executive course for C-level managers and a Bootcamp for professionals who will manage and work on the data sciences project. This combination is essential, Van Leeuwen explains. "After all, for data science projects to succeed, you need a purpose, business adoption, and skilled people. The first and the last part of the course is about business adoption. C-level managers get an overview of the Data Science landscape and learn about the strategic scope of Data Science. In the Bootcamp, professionals get trained in applying data science. They will learn how to set up a proof of concept from an initial idea to analysing data. "We will share best practices on scoping the project, preparing the data, and defining the model. Also, we share our deployment and user acceptance strategies to generate business value."

This mixture of business adoption and technical execution is unique, says Van Leeuwen. "Most other Data Science courses are focused on either technical depth or the strategic part, not on the combination. We are bridging the business adoption with the actual application of the data sciences so that the alignment between data science and business is secured and people learn about each other's field. When people leave the course, they know how to set up a data science project and how to harvest low hanging fruit."

Core to the course is the Team Data Science Process (TDSP) framework. This data science methodology from Microsoft provides a lifecycle to structure the development of data science projects. "TDSP", says Van Leeuwen, "is a nice flow of essential steps for a successful data science project. It starts with business understanding, acquiring and preparing the data, the modelling and ends with deployment and user acceptance."

Another distinctive element in the course is the individual coaching sessions, says Van Leeuwen. "People can bring in their own data to analyse. These could be sales data with client information or a product list with feature descriptions. The goal of bringing your data is to show how easy it is to get the first insights into a data set and bring these to the next level."

The Applied Data Science Bootcamp will be held online on 7, 8 and 11 June 2021 and on-site on 3, 4 and 5 November.