Advice From A Mechanical Engineer Considering Data Science

I had a recent conversation with a mechanical engineer that is considering a career change to data science. The MechE reached out to me after hearing my recent talk on the Super Data Science Podcast with Kirill Eremenko. It was a thought-provoking conversation, one that others can possibly benefit from. As discussed in the podcast, I was in this very situation.

Q&A

Q1. In your interview, you said you moved into management with "little background in data and that you started capturing and analyzing data to better understand the business and what to do. I'm in a similar role that straddles engineering and sales/business development in a machine manufacturing company (we use standard ERP and CRM systems). So, for someone in that role, what data do you recommend capturing to not only explore the data science field and grow professionally but also to prove the value of data science to management?

A1: Focus on the business KPI's. One might be reducing quality issues. Another might be increasing sales revenue. Use data to your advantage by gaining insights, and making decisions based on this data. The more you can codify this decision-making process, the better you will make decisions and the more value you will provide to the organization. Keep track of the value you provide. Ideally, this will be in terms of dollars/savings/increased revenue, etc. When management sees the results, they will be surprised/impressed and naturally push you to do more of it.

Q2. What tools do you recommend learning that would be most useful? (e.g., Excel's advanced functionality, Python, R, visualization, etc.)

A2. Excel is needed because everyone uses it, but it's not the best tool for the job. R is better for business because of its ability to make great, business-ready reports using Rmarkdown and knitting to PDF. When you become more advanced, use Shiny for interactive data products. You can build web apps others can use to implement your logic into their decision making.

Q3. What process do you recommend following to develop data science skills to be able to transition into a full-time data science role? (e.g., online learning, open source projects, blog to showcase projects, etc.)

A3. All of the above. A blog is good for your learning because it forces you to (1) know your material and (2) break through the barrier with most data scientists which are bashful about displaying their work. People hate criticism, but the productive criticism helps you grow. Learning R from the beginning is overly difficult now. I'll be putting some blog articles together in the coming weeks. First is "Six Reasons To Learn R For Business". Second, is "Data Science for Business: How to Learn R in 12 Weeks". I think these will help.

Q4. And lastly, how would you compare/contrast mechanical engineering with data science? Three big reasons why I'm interested in data science are 1. the technically challenging aspect of the work (disappointingly, I've found many MechE's roles aren't very analytical), 2. significant opportunities to work in a variety of fields (much more than a mechanical engineer, in my view), and 3. the open source, collaborative community where new technology is always being developed. If you have any comments or advice on this topic, I'd appreciate your insight (because ultimately my goal is to move towards a consulting role in data science, which is exactly the transition you've accomplished the past few years).

A4. I agree with you. Traditional mechanical engineering can be boring (with some exceptions in the areas of robotics). Data science is rated as the "sexiest job" for a reason - because it's evolving. This growth makes things interesting, and the changing landscape creates opportunities. It's a technically challenging field that you will likely enjoy, but more importantly, it provides real value to organizations that adopt a mindset of using data to drive decision making. Consulting is difficult, but you can start in your organization by learning new skills and showing management that you might be able to help even in areas that aren't that obvious. Don't be afraid to learn new tools. You won't regret it one bit. And, it will make you more valuable.

Wrapup

There you have it. A short Q&A that discusses some suggestions on how to navigate the challenges of transitioning into data science. The biggest point I want to make is: you can do it! It's not easy, but the journey is well worth it.

About Matt

Matt's the founder of Business Science, a consulting firm specializing in applying data science to business. Matt regularly contributes articles to the Business Science blog, a great resource for those looking to learn. He's also working on Business Science University, an online educational platform designed to teach novice data scientists how to implement advanced machine learning algorithms and build ML-powered web applications within their organization. Courses will be available in early 2018. You can learn more about Matt and Business Science by connecting on social media:

• LinkedIn: Matt // Business Science
• Twitter: Matt (@mdancho84) // Business Science (@bizscienc)

The biggest point about learning data science for business: You can do it!