Meg: What role has data science played in shaping your career trajectory? Was there a turning point that made you realize the importance – and potential – of data?
Gian-Franco: In my job as an application engineer, I realized quite quickly that collecting, handling and analyzing data is extremely important. At the beginning, that process was a bit time-consuming, but the more data I had access to, the better the outcome. I realized that I shouldn’t waste time analyzing data without checking it from different points of view.
In the semiconductor industry, our approach to analyzing process, cost, efficiency and impact requires a hell of a lot of data! And nowadays, research also requires a lot of information – technical and analytical results have to be addressed.
Meg: What kinds of objections do you run into most frequently when it comes to advocating analytical approaches? How have you overcome these kinds of challenges?
Gian-Franco: Most objections center around having too much data or too complicated, disparate data sets. This kind of complexity can result in losing a bit of value in the aggregate if it causes you to focus on things with less importance while overlooking more relevant or important factors. Up to a certain limit, I’ve changed the way I handle data, as well as my view on the value of data overall. The more data you have – even if it seems unimportant at the beginning – the clearer the picture or result will be in the end.
Meg: What are some common misconceptions you’ve encountered when it comes to industrial statistics and/or the value of analytics enablement?
Gian-Franco: People may think that some data is not so important at the beginning, but with a huge amount of data, sooner or later you can’t see the forest for all the trees! Collecting data – all kinds of data in fact – is becoming more imperative day by day in order to trigger, trace and analyze all types of value-adding activities and processes. In our industry, data is crucial to improving and stabilizing processes, getting pre-warnings or predictive trends before you run into trouble, reducing costs and gaining time.
Meg: How have you maximized the value you get from JMP? Is there any feature or aspect of the JMP community that you feel other people are underutilizing?
Gian-Franco: In my personal opinion, the power of JMP is not yet seen by technical people working in the area of support, problem analysis, and searching for fact- and data-based solutions. Even to just get a quick-and-dirty overview, a graphical approach will allow you to include or exclude parameters you’ve observed in your analysis. Then you can jump into a deep dive, problem solving for example by applying DOE.
Meg: What JMP resources, if any, have you tapped into? Anything you find especially helpful?
Gian-Franco: For me personally, the most helpful is the possibility of visualizing data. I always start to analyze my data with a distribution histogram followed by a multivariate analysis and 3D scatter diagram. This process gives me a very good idea of how to approach and/or solve the problem. Later, I step into DOE.
Meg: Do you have any advice for people who are trying to increase analytics adoption within their organization?
Gian-Franco: Check out and then distribute YouTube videos about JMP to convince others!
Meg: How about someone just starting out with JMP?
Gian-Franco: I would definitely recommend looking at the JMP software channel on YouTube. And simply try it for yourself, starting with a clear, well-arranged, small amount of data.
Meg: Is there anything you know now when it comes to analytics that you wish you had known years ago?
Gian-Franco: Most people start with Excel data sheets – two-dimensional graphics that can be well-arranged up to a certain threshold. Sooner or later, though, they’ll reach the limit of what can be done in an Excel sheet and start learning how to use a pivot table in Excel. That might help for a while, but sooner or later again they’ll end up with limitations. Then comes JMP, and they can take a huge step forward all at once.