2
$\begingroup$

Apparently, data science can be applied to the Aero industry, but how is a question that I still can't find an answer, and which proves to be incredibly elusive online.

I don't mean the Business Intelligence positions, I want to get more involved with the engineering team, working on the products themselves (airplanes, rockets and satellites). I think GNC (Guidance, Navigation and Controls, also called AOCS for sats) is interesting, but I don't want to limit myself to that. By "engineering" I mean working in the design phase of the product itself, rather than with post-market features (such as maintenance prediction).

Can a Data Scientist be useful in engineering, even without much domain knowledge?

Is there anyone familiar with this world that could provide some insight? Thank you

$\endgroup$
7
  • 1
    $\begingroup$ Or a FOQA analyst. (Somebody with the time to explain what that is can flesh this idea out into a full answer.) $\endgroup$
    – Ralph J
    Sep 10 '21 at 21:52
  • 2
    $\begingroup$ Can you please clarify what you mean by "engineering team"? At an airline? Navigation? Regulatory bodies? Aircraft designer? Manufacturing? The narrower the scope, the more answerable (within the site's limitations) it will be. $\endgroup$
    – ymb1
    Sep 10 '21 at 23:24
  • 2
    $\begingroup$ This is a very speculative question. I don't see what you can get here you can't get from a simple search, that is opinions. Data science, ex big data, ex analytics, ex business intelligence, ex statistics, is not always what is expected by data scientists. As of today this is often a buzzword for the customers. That said: A day-in-the-life of an Airbus data scientist $\endgroup$
    – mins
    Sep 11 '21 at 11:21
  • 1
    $\begingroup$ @temporario1001: You need to at least specify a product, again due to site limitations; as it is right now, it's a discussion question that may be unsuited to Stack Exchange. Here questions need a defined scope. $\endgroup$
    – ymb1
    Sep 11 '21 at 15:51
  • 1
    $\begingroup$ Voting to keep open - Not an opinion but a fact that aeroplane design requires availability of data. The question is particularly about aviation: aircraft-design is the label with the most hits. I understand the question as is, does not need any additional clarity for me. $\endgroup$
    – Koyovis
    Sep 13 '21 at 8:14
3
$\begingroup$

I am an aerospace engineer in a drone company and part of my job is "data science". We have a lot of flight logs on our server, which are uploaded automatically after a flight was done. I am using these as one part of my data source. Other sources can be simulations and specific component tests. As you can see there is a lot of data avaiable, even at our small company. With this data it is possible to improve the flight model (System Identification), improve the controller or create an AI e.g. for remaining flight distance estimation. Also finding out if the health of the Sensors degrade or the servos arent working properly is one part of this. Of course there is much more which can be done. But commonly we are using Data Science as a tool to improve our "aerospace"-functionalities. Data Science helps therefore to improve the understanding of our system.

At least at my job domain knowledge is very important. Knowing how to calculate Coeffcients or knowing which parameters influences e.g. Power Consumption is crucial.

$\endgroup$
4
  • $\begingroup$ Interesting to hear. My problem would be the lack of domain expertise though. Are there any Data Scientists that work with you that do not have background in Engineering? $\endgroup$ Sep 11 '21 at 15:44
  • 1
    $\begingroup$ @temporario1001, my questions back to you would be: Are there any data "scientists" who have do not have a background in anything except "data"? Generally scientists have a background in the specific discipline they are practicing science in, right? Wouldn't using the scientific method on data then just be one of the functions of doing research in that particular field of expertise? (rendering the researcher a "scientist"...) $\endgroup$ Sep 11 '21 at 21:42
  • $\begingroup$ @MichaelHall That's a very good point. I mean that I don't have a background in Aerospace Engineering, but in Physics, which is not directly applicable to AE jobs. So I guess the most useful thing I have to offer is Data Science skills. $\endgroup$ Sep 11 '21 at 22:03
  • 1
    $\begingroup$ @temporario1001, I worked with a fellow program manager in an aerospace electronics company who had a degree in physics. Sometimes you just need to get your foot in the door to eventually get to working at something that interests you. $\endgroup$ Sep 11 '21 at 22:20
3
$\begingroup$

I'll try to give you my perspective on working in aerospace product/systems development. First, there's no single answer that's all encompassing of the potential scope of jobs in the industry. In a company with a thousand engineers, there's probably 600 different job descriptions.

As an Aero engineer starting out, I had a view of where my career would go. The reality was pretty much something entirely different. The reality is that college taught me some basics, but what was important was how to learn, the scientific method, and how to solve problems. Engineering is just solving problems.

I've worked space systems, avionics, and weapon systems. After that, I moved into civil aviation avionics, mostly comm, nav, and surveillance. I knew development process, I had to learn the technology on the job. Along the way I was on occasion a manager.

As a manager I had to hire systems engineers. There were other managers hiring software engineers and EEs (digital design, RF design, power, etc.) and mechanical engineers (packaging). When you recruit, it's great to find 'the perfect fit' but they almost never exist. So you look for the skills necessary to do the job.

If you were to read the job postings, the would typically say: Systems engineer, must have a Bachelor's degree in engineering, physics, or mathematics from an ABET accredited university. And the it will talk about skills needed for the job (and a description of the job.) I should also point out there a number of jobs in the engineering team that don't have the title engineer. There are Analysts, Developers, Architects. A lot depends on HR decides to classify positions.

If it's a junior level position, under 5 years experience, I'm hiring based on what a coworker coined the Three "A"s - Attitude, Aptitude, Ability, in that order. Attitude is everything. I can't fix that. Aptitude is is important as they'll need to learn. Ability is good, but if they have the first two, they can learn.

Very few will have the perfect background. At higher experience levels work history is important, but I almost never had an exact match. Some did development in a non-aerospace business, others had the degree and some non-related aviation experience. So even here the 3 "A"s apply.

So the answer is read the job postings and figure out how your skills align with those listed and try fill the gaps. One quick way to gain a good basis in aviation is to take private pilot ground school. Actual flying lessons will help, but ground school will introduce you to the aviation world. And then go for it.

$\endgroup$
2
$\begingroup$

I think the dilemma in your situation is, that pretty much any engineer is a data scientist in some sense, as engineering deals with gathering information, analyzing it, making conclusions and then advancing design of doohickeys and whatchummacallems.

Airspace engineering requires a lot of specialized knowledge, and thus a sole training of data scientist most likely won't cut it. Data science as such is (according to my understanding) at its best when large sets of data or information need to be dealt with. And by large I mean stupid large.

I'm not sure if even the largest airspace entities have enough such projects to justify employing a data scientist, it may very well be a better option to subcontract this on project bases.

P.S. speculative answer, may get deleted, read quickly 😃

$\endgroup$
1
$\begingroup$

Yes data scientists have a place in aerospace design engineering, as aptly demonstrated by Torenbeek, Synthesis of Subsonic Airplane Design. A treasure of statistical design information. We went through a pre-design process at uni, using the book as a design tool.

enter image description here

For instance the table above, one of very many, lists design parameters for the initial phase of an airplane design. Data of existing planes will have to be explored for a new pre-design, not only geometrical data of course: fuel burn, range, payload, weight prediction, drag calculations etc are all initially determined from statistical data.

The table above was used in this answer, which contains a further example of data analysis from this report, Aircraft Drag Polar Estimation Based on a Stochastic Hierarchical Model. Yes, please read Fowler flaps for Flower flaps.

Data science has a very large role in aeroplane design. Without it, the design would be multiple times more difficult.

$\endgroup$
2
  • $\begingroup$ could you fix the second link? The one with the answer to another question? It is currently linked to the image of the table you provided. Also, this seems very interesting, but I am afraid it will require domain knowledge, which I don't have a lot of yet, and therefore it will be out of reach for "pure" Data Scientists. Am I correct? I have a background in Physics, which is not applicable to aviation. Thanks a lot for your reply though, it's really helpful. $\endgroup$ Sep 13 '21 at 13:42
  • $\begingroup$ Fixed the link, thanks. Yes obviously domain knowledge would be required, should not be a problem to obtain with a physics background. $\endgroup$
    – Koyovis
    Sep 13 '21 at 23:59

Not the answer you're looking for? Browse other questions tagged or ask your own question.