I wanted to share a very interesting blog post that Wilhelm Landerholm (@landerholm) wrote today on his Swedish blog, because I think it is worth a wider audience. Google Translate does a pretty good job of translating it (as Daniel Nicorici (@daniel2nico) pointed out on Twitter), but there were a few rough spots, so I produced a lightly edited version of it below with permission from the author. So now I give the word to Wilhelm – please enjoy! (For the record, the statement about robots which also occurs in the name of this blog post was written by Mattias Östmar (@mattiasostmar) on Twitter before appearing in the blog post.)
Posted on January 21, 2015
This wonderful future, and this constant discussion of what’s to come. There any many prophets, and many who say, in response to events, “What did I tell you?” Naturally without mentioning all those times they were wrong. What will that prophet that recommends you to fix the interest rate at 3.78 percent today say in the future?
In my world, you build models to explain future events. It is therefore always a bit funny to brainstorm about the future of data analysis and the ever so popular concept of Big Data, because what is described is often actually the world we live in today. As an example, today I saw a slide that began with the words, “Those who are best at Big Data are those who will be the best at operations”. But that is not the future – it is a description of how it is today.
I’ve built so many models I would have lost count if I hadn’t documented them. Several of them have been the difference between survival or death of a business. These results achieved by myself and my colleagues have also led to my profession being described as a sexy future profession. I hesitate to concur. Today, no one within my area talks about “creating analysis teams” within the company, even if this is what the Swedish companies are currently doing. Those in my world are busy seeking interfaces between my day-to-day reality and customers. The focus is on automation, on AI and Machine Learning. I am one of those who believe that the future does not have room for the Data Scientist of today. The future belongs to those who can control “robots”. And I am not talking about robots like R2D2, but robots like IBM’s Watson.
Today, I build models for large corporations. For companies where the effects of my work can be enormous. But in the future, these solutions will not restricted to a few large corporations. They will be a part of everything we do. The Internet of Things will generate such volumes of data that analysis must be moved from BIG to MICRO. This will, in turn, mean that the analysis must be automated. This is where we work on the future today. Today I am putting together “self-driving companies” in a Raspberry Pi and a smartphone. In a few years, we will see self-driving cars. However, I am not sure that there will be a market for everything that moves, and that’s why those of us who were early to jump on the Big Data and the Internet of Things train are now looking for contacts with those who own demand.
No matter how many people out there are shouting that the company must build its own analysis team, a BIG DATA UNIT – you’re ten years behind. Because the work that I have done for several customers can now be replaced with a computer for 500 SEK [about 60 USD] at Kjell & Co [a common Swedish chain store for electronic peripherals] and a few [predictive?] models. I focus on being the supplier of the models. The future belongs to those who can automate the analysis. For those who own the robots. Along with those who own demand, they will change, to a large extent, how business is done.