Data helps increase profits in big companies such as Apple — Interview with Apple engineer Julen Aristoy.
Data is everything in today’s world, and even when you think something is not using data, it is. Engineers for companies such as Apple use data to make a product better and fix problems they may have. I talked with Julen who is a display quality engineer at Apple and he explained to us how data is used by his team, how they process it, and how important it is to visualize. The reality is that without having a way to visualize the data, data will be a lot of information without a purpose. For this reason, Julen is explaining to us a little bit of what his work is about but before listening to this conversation, let’s look at this box plot below. This is a box plot Julen sent me and he explained this is a comparison between four manufacturers and the red line is where you don’t want to be, making us understand the purple or D box is the worst manufacturer.
…
Let’s listen here:
Maria: Hi guys, My name is Maria Chacon and today we’re going to talk about data. For this interview I am here with Julen. Julen is an engineer and quality control at Apple in San Francisco and he is going to talk about his job and how it connects with data.
So Julen, thank you so much for being here. Please tell us about yourself and your role at Apple.
Julen: Hi everyone, this is Julen. I am a display quality engineer at Apple and I graduated three years ago from a material science degree and so currently, I’m working centrally in the Display area of the apple engineering team.
Maria: So how is data used in your job?
Julen: We need to use data in the job since every single part of our products has different failure modes. So we need to be able to detect those failure modes and see how in control they are with respect to the manufacturing of each module line.
Maria: And having in mind what you just said about data and your role. Do you think data is important for your career and why?
Julen: Data is very important for my career, essentially, because if you’re an engineer, you need to be able to draw conclusions from different set pieces of data. So for example, if I pull the raw data from a certain production line, I need to be able to tell if it stays within spec, and how capable this manufacturing plant is of doing this process.
Maria: How do you work and process data?
Julen: We must be able to basically analyze large data sets because different vendors do different parts of the display production line and we must be able to tell how well certain vendors perform in comparison to the competitors. So essentially, we set the spec lines, and then we need to be able to tell how much the yield of our data set and as well as how well the, basically the performance capability of the process is and in order to do that, we essentially, usually check the SPC charts, or statistical process control charts and then from there, we can basically draw most of our conclusions. If I need to handle small datasets or medium sized datasets. Usually I use a program called Jmp. This program allows me to basically plot the data in a way that I’m able to visualize it easily. And as well as basically manually manipulate and calculate maybe the yield CPK, or the variance or other parameters that we need to basically check if the data is working properly or not.
Maria: And how do you think your job will be different if you didn’t work or didn’t have this data to work?
Julen: If I didn’t have the capability of checking the data, my job would be substantially more difficult because without data, it’s much harder to basically verify, verify how a vendor is performing comparison to the others. So we would only be able to tell by the final and product yield and this would basically waste a lot of money in the production line and then in the end, we would essentially would not be as profitable as we are currently with basically analyzing the difference datasets that we get downstream the production line
Maria: And talking about how you work with data, how do you collect data and information?
Julen: In our job, we don’t directly collect data, we basically have to ask the vendors to upload or share the data to us and then from there, we’re basically able to analyze it. So as a more indirect way of doing it. But at the same time, it’s also slightly harder because of course, we need to keep the data integrity all the time in check. So if we see any weird variance or any problems, we need to basically ask the vendor and hold them accountable.
Maria: How important is it for you to visualize this data?
Julen: Visualizing data, as I mentioned before, is critical. Since if we only have a large basically either Excel sheet or just overall data sheet, we wouldn’t be able to tell if something is going wrong. So that’s why I use Jmp and maybe other programs to basically plot the data, and maybe check the SPC charts, check the correlation between two different parameters, and other things that you could check.
Maria: For you, what do you think is the most interesting thing about your job?
Julen: For me personally, my job, since it’s not 100%, data driven, I would say that the technical part of understanding the different processes, and basically the physics behind the different components is what basically drives me most days. But of course, it is important for me to understand the data analysis part because without that, basically, I would only be technically driven. I wouldn’t have any hands on or more process experience. So that’s why I think data is critical for my role.
…
