Cover image for How To Build Your Data Science Career Change Action Plan

How To Build Your Data Science Career Change Action Plan

alanhylands profile image Alan Hylands Updated on ・8 min read

Your Job's A Joke. You're Broke. Your Love Life's DOA.

You've spent the last ten years grinding it out at a job that you don't exactly swoon over when the alarm clock goes off. It's starting to feel like life is passing you by and you know there must be something out there that you can really get your teeth into. But which direction do you go in to find it?

So you read some articles online about Data Science being the sexiest new career path since the last sexiest new career path. And it starts to sound interesting. Who wouldn't want to find out more and see if this whole data world could open up for them?

I'll be honest now. Coming into the data field at age 30 or above without any prior experience or tech background can be tricky. To short-circuit the whole "how do I even get started?" problem I always suggest planning backwards.

What is Backwards Planning?

Backwards Planning is a great concept I learned from Amy Hoy and Alex Hillman from stackingthebricks.com. Essentially it means looking at what you want to get to as an end product (e.g. writing a book, building a data product...hooking a data analyst job) and working back from there. What does the end result look like? What’s essential and what’s nice to have? What do I need to prepare up front? What has to come first? Etc. etc. All questions to take you backwards from your end goal to where you are now so you can fill in the blanks in between. (You can read more about using Backwards Planning in Amy and Alex’s book Just Fucking Ship).

When it comes to approaching a career change you need to look at getting that new data analyst job as your ultimate end goal. If you know your end goal then you can plan backwards to find the steps you need to take between there and now to get you to that place.

These five steps in fact.

The 5 Step Guide To Backwards Planning Your Career Change.


Go to a jobs site like Linkedin or Indeed and search for "junior data analyst" jobs in your hometown. Go through the wordy job descriptions and make notes on what they are asking for. Pay particular attention to the technical skills, education and soft skills sections.

When you've got enough data together, you can start to pick out the actual skills and requirements that will get you in the door at more of these jobs. Got those skills already? Great, you’re one step ahead. If not then it’s time to work out how to get them.


I'll go out on a limb here but you will probably see Excel and SQL pop up most often in Required Skills. These are the fundamentals of working in a BI or analytical role in the vast majority of companies, both large and small. If you don't have these skills then you know where to start your learning. You don't have to be the world's best Excel jockey or SQL superstar either. You'd be surprised how many people get tripped up at the first hurdle on these entry requirements. Make sure that isn't you.

Excel and SQL really are two givens in terms of necessary knowledge in the broadest range of companies so it makes sense to pick those off first. At this level there's no sense in spending a fortune on training when you have the likes of Udemy courses available for under 20 dollars. Have a look on there, check out the class ratings and reviews and jump on in. I’m a big believer in getting the Pareto Principle and making sure you get the basics right. If you can nail 80% of the most important SQL commands or Excel functions you’ll be streets ahead of the entry level competition. Even those graduates from fancy colleges. In fact ESPECIALLY those graduates from fancy colleges.

You can get into a coding language like Python or R afterwards. Having a curiosity to dig into datasets, thinking of questions and then finding the answers from the data is far more important than listing 20 different programming languages at this stage of a career change. Those of us who scan CVs and resumes when hiring analysts can tell immediately that you are a bullshit artist if you rhyme off a shopping list of programming languages and database technologies. Especially if you are straight out of university. Better to know the important ones well.


Do you need a 2 year college course or bootcamp in Analytics to learn Excel and SQL? No. Not at all. Despite that, not having even a bachelor's degree will tend to be a barrier to getting in the door of many organisations. I don't agree that they SHOULD require it. I've worked with plenty of excellent folks who don't have them and it hasn't stopped them being damn fine analysts but HR departments often disagree. If you don't have one, search out the jobs that don't ask for it and see what they want instead. Or just apply anyway. Let them weed you out if they want but don't rule yourself out unnecessarily.

In fact, that’s probably the most important piece of advice I can give for your job search. Companies employ masses of HR people who are only fit to scan for buzzwords on application forms. If they have a strict policy like only hiring graduates for entry level analyst jobs or only hiring from top level colleges then it’s their loss. Angela Bassa is the Director of Data Science at iRobot. She covered this very subject in a Tweet thread about how she wouldn't have passed the Master's or higher graduate-level degree bar herself on a team that SHE set up. And she went to MIT! It's a ridiculous example of artificial gatekeeping. But in a crowded job market it happens.

Some companies are a little more forward thinking however. They give their recruiting managers a little more leeway when it comes to who they bring in for an interview. As more people change careers in their 30s, 40s and 50s than ever before, it makes sense to not rule out picking up a diamond that may have otherwise been discarded. Don’t limit yourself by deciding you aren’t fit to apply. That’s someone else’s job, not yours. Fortune favours the brave. If that’s you then the next step will be where you can really shine.


Let’s say you've worked as a chef. If there is a more highly stressful, on-demand, pressurised working environment I've yet to see it. You should have learned plenty of skills running your kitchen that can transfer over to a BI analyst role. See what the job listings ask for (e.g. "Ability to be self-managed, work independently as well as to collaborate within an agile, fast-paced, dynamic team environment.").

Ever worked with other people? Have you worked on something on your own? Been under pressure on the job? How did you deal with it? Was there something broken in the process you were meant to follow? How did you fix it?

These are the kind of open ended questions that really show how you will deal with situations in the real world. It's not all about the technical chops especially for folks with some previous life experience in other arenas. You mightn’t be able to quote the exact dictionary definition of what a SQL window function is but if you can show me you know how to Google things, I’ll respect that more.

Prepare some examples of when you have displayed these soft skills in your career and life to date. Don't be afraid to use non-work related examples either. I've had people use their experience in sports clubs or other organisations to display their soft skills at interview. I'm always happy to explore those as it shows they can transfer the skills beyond just the office work environment.

(I should say that I hate the connotations of these highly valuable skills being termed “soft skills”. It’s as if they are an afterthought or not quite as useful as the (presumably) “hard” technical skills of programming and working with databases. Anyone who has ever had to manage a team of other grown-up human beings will know this is horse shit.)


When you've built up your skills a little, it's time to show people what you can do. Write blog posts documenting how you approached a project. There is a common misconception that seeing as everything in the world has already been written about then there is no point you doing the same thing. That’s bullshit. Write about what you have learned regardless of how basic you think it is. Get used to communicating your process and what you’ve found. HINT: these are REALLY important skills to have as a professional data analyst.

Or get a dataset, analyze it and present what you've found. (Just don’t use pie charts. Please). Interested in sports? You'll find plenty of sports data online, jump right in. Got a book on your coffee table full of the US Government’s notes on Italian-American gangsters of the mid-20th century? Source the data, cleanse it, wrangle it and analyze it (just like a project I’m currently in the middle of writing up).

For your online portfolio, Google and read as much as you can from other data science folks who have posted theirs online. Unfortunately a lot of us aren't able to discuss actual projects we've done on the job due to NDAs etc. but there are still a lot of great resources out there to pick over. See what the format is like. How did their voice come over in the writing? How did they present it? Github/Jupyter Notebook/R Markdown? It’s all research, all analysis. Even the searching and researching is building your data skills. Learn from everyone, pick the best elements and plunder all you can to learn more.

The Final Pep Talk.

It might look like a lot of work but this is a great field to get into. Career progression through the many different areas of data science is good and well compensated. Getting a foot in the door is the hard part but it's well worth it and there is no better time to start than today.

Overall, I'd say just work hard on picking up just enough new skills to be able to demonstrate you can use them. Don't want until you consider yourself an expert to get applying for jobs. The real learning, as in any field, is done on the job as you are adapting to real world situations. And, in the immortal words of The Carpenters, on that front “we’ve only just begun…”


If you enjoyed this article you should sign up for my weekly newsletter at SimpleAnalytical.com. You’ll get advance notice of new articles, along with a veritable treasure trove of added data-related bonuses. Cheatsheets. Ebooks. Code samples. And more.

Posted on by:

alanhylands profile

Alan Hylands


Solver of difficult data problems. Builder of online things. And offline things.


markdown guide

Interesting, also the fact that there's no mention of the common clichéd high tech things like AI or Machine Learning (or even of math & statistics) ... so my takeaway of this is, just learn the basics and know what your "soft skills" are, get confident, and apply ... SQL and Excel, it can't get more down-to-earth than that, love it!


I'm a great proponent of the "data science is a VERY broad church" school of thought. ML and AI are just a small sliver of the areas you might get involved in but they certainly aren't the whole thing.

I know there is a lot of gatekeeping even around the phrase "data science" and it annoys me a lot. We should be encouraging as many people to get into all areas of analytics/data science as possible.

A lower barrier to entry path is building SQL and Excel skills to help encourage analytical thinking as a stepping stone to further technology development as they progress.

I've probably undercooked the really important part which is data literacy. Get that sorted early and the coding/tech parts can be bolted on. Without data literacy and business knowledge, you can know all the programming languages in the world and still be virtually useless as an analyst.


Interesting concept, Data Literacy en.wikipedia.org/wiki/Data_literacy ... I think that's something which even kids in primary school could be successfully exposed to.

I'm 100% with you on the gatekeeping thing, that's just an elitist trick to keep "unwanted outsiders" out of one's privileged niche, keeping salaries high and promoting an air of mystery and "high tech" around the subject ... not good for companies, and also not for motivated and potentially talented candidates.

Virtual high fives on both points.

We have to take the long term view on embedding data literacy so primary school is the perfect starting point. Make it second nature for all kids to understand the basic concepts and reap the benefits in all spheres in the future. Same with personal financial literacy but that's a hill to die on another day.

Right! same with computer literacy (but that's commonly being taught already), and "social media" literacy, and I can go on like that. World is changing, kids need to be prepared.


Thanks for sharing. Data Science had really piqued my interest lately. But what should I do if it hasn't been my day, my week, my month, or even my year?


Don't worry John. (Clears throat nervously) I'll be there for you.


Great to hear you're looking to pivot your career ...


:-D If I had a penny for every time my wife managed to use this quote in real-life, I'd be richer than the cast of Friends.


Hi Alan,
I have a project question for you. You may consider it out of context but let me give a shot.

I found a paper of the Top 10 Algorithms for DM. I loved the idea so I am striving to put a brief booklet together on what these 10 Algos can do given a set of Biological data. I would like to show R code, output then describe the benefits and downfalls for my particular dataset and tease them all apart. Do you think that would be well received?


That sounds spot on Matt. It gives you room to showcase your technical chops with R and your subject matter knowledge from your Biochemistry background. A lot of people get caught up on just the tech side of it and forget about making the analysis relevant to a real-world use case. If you can take the raw data, wrangle it, present it and give proper analysis on what you've found, it would be very well received - especially when it comes to making your CV/resume stand out from the rest of the pile.


Great, thanks for the encouragement,

Are you familiar with the R package 'caret'? Powerful package, it actually allows one to run and compare multiple methods with the same syntax, basically the same command(s)...

BTW, I had an high school exchange student friend from Carrickfergus. Do you know it? ;))

Not familiar with the package but I'll put it on my list to check out.

And I do indeed know Carrickfergus. They've got a great old Norman castle up there and a colleague from my $DAYJOB is from the town. If you are into crime fiction, the Northern Irish writer Adrian McKinty sets his Troubles-based Sean Duffy detective series in Carrick (where he himself grew up). Small world!

The comedian Steven Wright says, "small world but i wouldnt want to paint it."