Roadmap for Transitioning to Data Analytics

If you are working in another field, how can you transition to data analytics?

You may have a university degree in an unrelated field, or have been working in a completely different domain. You may be interested in transitioning into a data analysis role for a while, but do not know where to start. If this sounds like you, keep on reading! 😉

Two ways to get into data analytics

Essentially there are 2 ways to get into data analytics:

• (1) Completely self-taught: then cleverly combine analytics skills with your current domain knowledge either from your earlier study or jobs to gain competitive advantage;
• (2) Taking a data analytics degree or bootcamp.

In this blog post we will focus mostly on the self-taught route with the target of becoming a data analyst, which is a good starting point to a lot of other data science roles. This is also how I started 6 years ago, back then I was excited about healthcare analytics so this is the area I started out with.

In this article we will talk about what you should learn and prepare in terms of:

• Required skills
• Portfolio projects
• How to approach recruiters

Things are changing rapidly, with many concerns around the potential of AI to displace jobs and how that may affect you when pursuing data analytics. I will also share some insights on this at the end of this article.

You can also watch the video version of this blog post below 👇

So what does a data analyst ACTUALLY do?

There are actually a lot of different job roles that involve working with data and turn it into insights. Long ago, people who did this kind of job would be called statisticians or actuaries if they work in insurance industry.

Nowadays, we are sometimes too hung up on titles and it can get confusing real fast. If we look at job postings we see Business Analyst, Research analyst, Analytics consultant, Junior Data scientist and so on. Not to mention Data analytics + X kinds of jobs, that are roles that are specific to a domain, such as HR data analyst, sales analyst, marketing analyst, etc.

I find it more useful to understand the nature of the job and what skills you actually use. So in this article I simplify all these titles as data analyst.

A data analyst is someone who analyzes data.

While Data analysis is a process of inspecting, cleansing, transforming, and modeling data with the goal of discovering useful information, informing conclusions, and supporting decision-making.

For this reason data analysts often do a large chunk of a data scientist job, except for the predictive analysis, machine learning and software engineering which data scientists may focus more on. So working as a data analyst can give a great foundation for you to progress into a data scientist, data engineer and other data science job roles.

Job prospects

The median salary of a data analyst is $95,000 in the US. The demand keeps rising for people who have domain knowledge and at the same time know how to analyze data.

As a data analyst, you’ll have the opportunity to work in many different domains like healthcare, finance, banking, logistics, human resources. You can work for startups, agencies, corporates or, consultancy firms like I do. You can work full-time or freelance basis, or even remotely. So you have a lot of flexibility in your career!

Required skills

When it comes to the most important skills for a data analyst, let me show you this skill model to help you visualize it.

(1) Basic math and statistics:

On the most fundamental level you need some basic math and statistics. Just basic high school or undergraduate math and statistics is sufficient, unless you want to do more advanced analyses or machine learning in your job. Most of the time as a data analyst you will do some descriptive statistics on the data, like calculating average, min, max, median, standard deviation etc. Depending on the projects, you might also do outlier detection, hypothesis testing, linear regression, clustering and even machine learning. It depends on your job and how technical it is. I would highly recommend you read this classic book “How to lie with statistics” by Darrell Huff. It’s a really fun book to read, pointing out the most common mistakes we make when drawing conclusions from data. I think everyone who works with data should definitely read it.

(2) Technical skills:

Secondly, a data analyst needs tools, i.e. specific technologies or software. As a data analyst, you want to get 3 things done almost on a daily basis:

• Extracting data
• Analyzing data
• Visualizing data to communicate insights and tell a story

There is a variety of tools that can help you with each of these tasks, or all of them. The most in-demand skills are SQL, Excel, Tableau, Python and R. You can choose to learn PowerBI as well. But if you know Tableau, it should be quick to learn PowerBI and vice versa. In the future, we can expect to work with big data as a data analyst. So it may be useful for you to learn the big data versions of these tools as well, for example SparkSQL for querying big data, and PySpark for working with big data in a Pythonist style.

(3) “Soft” skills:

On the next level, we have the soft-skills. It is critical for a data analyst to be able to communicate insights, as you are going to be working with business stakeholders, product managers and other key stakeholders. You will need to ask questions and understand the business problem, and then iterate through different analyses and results with them, and finally come up with the some insights and recommendations for the business. Most of the times, communication and storytelling takes place in form of presentation, or documentation, emails, dashboard and so on.

(4) Domain knowledge

And finally on the top level, we need domain knowledge. It gives you the ability to really understand the business questions and also better understand the data. For example if you’re working in airlines industry, you will need to learn about how airlines work. If you do data analyses in healthcare it is also critical to know about the healthcare domain. This domain knowledge you can get from either your education background, or from working experience in the field.

How to learn the skills — Where to start?

There are many ways to learn the required skills for data analytics roles. Here are some tips:

• Regarding math and statistics, it can be quite easy to go into a rabbit hole and never feel like you have learned enough. For me, even after 6 years I still feel the same way. But it is okay to have basic to intermediate understanding. A lot of things you can learn on the job, so don’t be too worried about not knowing enough from the start.
• Regarding Excel, I would encourage you to really master it. You want to know the basic things like the back of your hand, such as VLOOKUP, XLOOKUP, Index match, conditional formatting, Pivot tables, etc.. You can go for more advanced tools like Macros and VBA to automate repetative tasks, and optionally PowerQuery if you need to connect different data sources. In a near future, a lot of the small tasks can be automated by AI. So the main competency we need is knowing how things work. It is much easier to ask ChatGPT to write some VBA code for you to fill an Excel template, if you already know how VBA works!
• Use transfer learning when learning coding: If you already know a programming language, for example SQL or R, it can be easy to transfer that skill to a new language like Python. When I first learned Python, I often tried to relate things from R which I was more familiar with. For example if I want to concatenate 2 dataframes I would google “how to row bind data frames in Python” because that’s how I would describe it in R. In addition, all the basics like basic concepts like variables, data types, functions, lexical scoping are almost exactly the same. So overtime you will develop the intuition for how to do things and problem-solve, in whatever language. The rest is simply to practice!
• Regarding data visualization, it is perhaps one of the most fun things to do as a data anlyst. Data visualization can be stand alone graphs that you can create in R, Python or Excel, or in the form of dashboards or presentations. For starters there are some good books such as Storytelling with data. For dashboarding, you can learn pretty quickly with some courses on Tableau or PowerBI, which are drag-and-drop tools with no coding required.
• Regarding soft skills like storytelling, asking questions and presenting, you will develop over time. But you can also proactively learn to become better at communication and storytelling by practicing it, by writing writing blogs about your projects, pitching the ideas and insights to your friends and family and see what they think.

Prepare a data portfolio

In a sea of candidate, you definitely need to stand out. And I believe a good personal portfolio is going to help you best showcase your skills and get the attention of recruiters. Doing projects is also a great way to put your skills into test. It gives you a sense of esteem and confidence, which is probably the most important aspect of it.

If you don’t know what projects to do, I have a few project tutorials on my channel, which you can use and adapt to your specific use case or domain area.

Here are some extra tips:

• Be sure you have 3–5 projects in your portfolio. I think 3–5 projects is a sweet spot for you to showcase different aspects of your skills.
• Do project about something you care about, find or create a dataset you are actually interested in, and put some LOVE ❤️ into it! Your passion will show! Looking at two projects, people can easily compare which one the author put more thoughts and commitment into them.
• Everything is hard, until you figure it out! Sometimes it takes a lot of courage to get over the mental block to get started or continue. A few months ago I started bouldering. Climbing teaches me a lot of lessons about getting over my fear, to fully commit to a difficult move. It is refreshing to know that I can train myself to think: I’m scared, but I’m going to do it anyway.

Approaching recruiters

Offline & online networks

After having some skills and portfolio projects under your belt, it’s time to get yourself out there. I find that offline network like friends, relatives, former classmates or colleagues is the easiest way to find jobs, this is also what worked for me before. But there’s also a lot of opportunities online.

LinkedIn is also a good place to reach out to people already in the field and recruiters. I have never actually tried to reach out to recruiters on LinkedIn, perhaps because 7–8 years ago it was not common to do this. But if you give it a shot, just be sincere and polite in your message. If you have any kind of connection to the company, even the slightest like you attended an event by the company or talked to someone there, make sure to mention it in your messages!

You may also want to double-check your grammar using Grammarly or let a friend proof-read it for you. Small mistakes can lead to a bad impression, especially when this is the only impression you get to make!

Don’t under-sell yourself!

From my experience, it is a common mistake to under-sell yourself, if you are really new in the field you maybe tempted to think “I just graduated and am still learning”. It may be true, but that is not what employers are looking for. You may want to frame it in a more assertive way, along the line of what you have done and what you could contribute. For reference also make sure to send the recruiter an up-to-date resume, this should include links to your portfolio projects like Github repos, portfolio website or other mentioning of your projects online.

If you have previous working experience or some “domain knowlede”, you should definitely leverage this. So even if you worked as a barista or cashere at a supermarket, this counts as your domain experience. Data analysis is nothing without context, without business problems. If I were you, I’d mention these experiences and try to link them with the job I’m applying for, try to tell a little story and show that what I learn from those experiences may benefit the company. It’s not always easy and straightforward, but I think it’s a very effective, clever strategy.

How ChatGPT will affect data analyst jobs?

Okay, talking about AI. There’s no doubt that LLMs and AI tools will transform analytical jobs at some level. A recent analysis found that large language models such as GPT could have some effect on 80% of the US workforce. Some creative and high-paying jobs that are most vulnerable are writers, web and digital designers, together with financial quantitative analysts, and blockchain engineers.

But I think the upsides are going to outweigh the downsides. A recent study shows that the AI tool helped the least skilled and accomplished workers the most, decreasing the performance gap between employees. In other words, the poor data analysts will get much better; and the good analysts will simply get a little faster!

As you probably already know, data pre-processing accounts forperhaps 70–80% of the work for many data analysts and data scientists. Most of it is pretty boring and tedious. I think tools like ChatGPT and GPT4 model can help us automate a lot of those tasks, simplifying and streamlining the data analysis process. Some of the best use cases for AI tools like ChatGPT for data analysis include:

· It can Create code for constructing databases, doing simple data cleaning, explorative data analysis and creating various charts

. Optimize your code and add comments and explanation for documentation purposes.

· Give you suggestions on how to communicate information to different audiences — executives, departmental heads, managers, and so on.

· Suggest data sources, for example “Where can I find data on financial fraud in the Netherlands?”

· You can also ask it to create synthetic data for training models or testing algorithms.

· Provide general advice on regulations, legal processes based on open sources on the Internet.

That being said, there are certain things generative AI cannot do, thankfully, at least not yet! Even today’s most sophisticated LLMs still lack key abilities like critical thinking, strategic planning, and complex problem-solving. Nowadays, it is not about simply doing the same thing cheaper anymore. Companies need to innovate and think outside the box if they want to stay in business. So as a data analyst, you have the power to help them do just that. LLMs are also not able to come up with jokes just yet, and being funny and likable is surprisingly important in building trust and relationships. So we can expect all these aspects are where human data analysts can shine.

The key for us is to keep on learning, acquire good fundamental skills, stay up to date on how to utilize AI tools and techniques to be better at your work. And also more than ever, we want to focus more on soft skills like empathy, communication and relationship building.

Conclusions

I hope this blog post gives you a better understanding and a roadmap of how to transition from another field to become a data analyst. Good luck with the next steps in your career! 😊🍀