Data Science in Transformation
Have you seen vanilla sky lately? Sometimes it is better to stop and think what has changed and whether I should start doing something differently. In cloud business, the last two years have been a transformation period, especially in the data science playground. Digitalization, big data, open source tools, modern cloud & analytics integration, advanced analytics lifecycle, outstanding visualizations, and value chain transformation, just to name a few fundamental shifts. Let’s open up a bit some essential data science topics: people, digitalization leadership, and tool & roles.
It is all about people
You see “Full Stack Developers wanted” everywhere. The same happens in data science. “Data Analyst” starts to be too narrow, “Data Scientist” is way too big and not always well understood. The best data science professionals master the full analytics stack, and the required skills are:
- dig up the right data
- utilize modern tools designed for big data
- R or Python, SQL
- analysis, statistics, machine learning, data intelligence
- creative visualization and communication
- integrate to smart apps and dashboards
- think like a data scientist
So what is the point? The same was done before. Now it’s about efficiency and value. Earlier, there were several people doing the above; now it can be done in optimal case by one or two individuals. Do we need superhuman employees? No, the new tools are helping a lot. Several weeks’ worth of advanced analysis can be done in two days. And even today, the creative human touch remains: pick the right visualizations, make it truly visible, and create the usability that makes people want to be data driven.
It is all about digitalization leadership
Company leaders need to have basic data science skills. Oh my, more learning? No, it is an opportunity instead. It is about the courage to jump in and ask the data. Today’s solutions provide easy ways to analyze, drill down, make comparisons, and so forth. They are really easy and cool to use! Advanced analytics requests can be completed in hours when you have a good architecture in place. The most important thing is that the analytics solution needs to provide immediate business value. Not once or twice, but constantly. Sounds too perfect? Maybe, but it makes a huge difference for business if you can use fact-based data in decisions 20x more often, right? Vanilla sky, stop and think about this a bit.
Tools & roles have changed a lot
It is not just cloud anymore but an integrated service; tools now talk to each other. The advanced analytics process has been implemented on major platforms such as Azure and AWS. Earlier, 80% of analytics work was manual and only 20% automated. Now this starts to be vice versa. Another example: Not that long ago, Forbes reported that analysts spend 60% of their time cleaning and organizing data. And they hate it; it must be done, but what a waste of time! This is a fundamental shift in technology enablers and every company should understand how to change themselves. Get rid of silos, get rid of ETL work, get over data modeling, make data truly visible, stop saying “but the data quality is poor”.
What does all of this mean?
The value of data has been increasing exponentially. Extracting the value has become ten times easier. Combining data in ways earlier considered impossible is opening new business opportunities. Fast, outstanding, and surprising results are discovered. Data science is fun again!
Thanks for reading, the transformation continues…