At the German SAS Forum, Sascha Schubert and I ran some hands-on sessions or workshops on data science and analytics. You might say that this was nothing new, but I noticed some changes from previous forums that I think may point towards wider changes in the analytical landscape. Here, then, are my key takeaways from the workshops, in the hope that they may be useful learning for others too.
Analytics is moving into the mainstream
There were two particularly interesting – and surprising – elements of our workshops. First, they were sold out, which is always gratifying on a personal level, but I think also indicates other factors at play. Second, the attendees were not just data scientists. There were also a number of business users attending, which is a change from our previous experience. I think these two elements suggest that analytics is now very much not just for nerds, but for everyone, right across organisations. This is consistent with the rise in the number of “citizen data scientists,” those who do some analytics, without analytics being a key part of their role. It may also be a signal of the keen interest that many business users are now developing in the insights that analytics can offer, and the fact that data-driven decision making is becoming the norm in many places.
Visual analytics is an ideal entry point for the ‘analytical rookie’
We demonstrated several visual analytics packages and options during the workshops, including visual data mining and machine learning. It became clear that these visual analytics options are much more appealing for people without a background in analytics, including those who have never used any kind of analytics before. These packages are, in fact, an ideal entry point, because they make it easier to “see” insights, without having to do a lot of digging into columns and columns of data. This is important for any senior team thinking about rolling out analytics in an organisation. You have to make it easy for people if you want to encourage data-driven decision making, and visual packages may be the best way to do so.
It is also important to ‘keep it simple, silly’ when talking analytics and insights
Visual analytics is chiefly appealing for the way it makes insights more obvious: It is easy to understand and explore. This provides an important lesson for analytics more generally: It is important to keep things simple. It can be challenging to convey analytical insights in simple language and without using jargon. It is particularly difficult when you understand the concepts very clearly. Analysts and data scientists, however, need to rise to this challenge and remember that business users and senior managers do not want to be “blinded with science.” Instead, they want to understand insights rapidly and clearly so they can explain the basis for decisions to others. Keep the language simple, and you will get your insight across much better – and your insights are also more likely to influence decision making as a result.
A single, unified web interface, serving multiple purposes, is hugely appealing
We have been talking about analytics platforms for some time, but there is definitely a step-change in requirements for analytics now. The huge increase in analytical need is driving increased interest in platforms, and it was clear that the participants at our workshops found the idea of a single, unified web interface, serving multiple purposes, enormously appealing. Customers seemed almost stunned at the idea of being able to do so many things in one interface that effectively provides end-to-end analytics.
Do you need to gain stakeholder buy-in, raise investment, and prove AI´s value to your organization? Download our comprehensive step-by-step guide and strengthen your efforts.
Everyone ‘still’ likes live software demos
Finally, something of a process point. Talking after the sessions, we got the impression that part of the appeal (apart from us, of course) was seeing software used live. Static presentations are all very well, but they do not really bring analytics alive in the same way as a demo, and they certainly don’t allow you to show the fun of data science. Everyone, from data scientists through business analysts to IT folk and managers, is interested in seeing the software in action. Of course, there are risks to demos – technical hitches and glitches are an inevitable hazard – but they are significantly better than a slide-only presentation, both for the presenters and the audience.
