How to become a Data Analyst?

There is no better way to start this article rather than with some good old data.

Photo by “Franki Chamaki” on Unsplash

Job posting for Data Analysts grew 16% from 2015 to 2020, according to the IBM report "How The Demand For Data Science Skills Is Disrupting The Job Market".

The demand is growing and yet those professionals are hard to find. According to the same report, those positions take 5 days longer than average to find a potential match.

It is still clear this is the sexiest job for the 21st century and if you are here it's because you want to find out how to get one.

The steps are quite straightforward:

1. Acquire the needed skills
2. Work with some real life data to get experience
3. Apply (and get) a job

Let's get started

Acquire the needed skills

While interviewing for an Analyst position I usually ask the candidate: “What are the skills that compose a good Data Analyst?”

We have the most varied answers. Some focus only on the hard skills, others at the soft ones, some get creative and mention things like “ability to learn”, “work as a team”, “know the impact of your data”. But the truth is:

You need a combination of soft and hard skills to become a Data Analyst. Most of the time, soft skills is what will make you excel

Let's deep dive on these two skill sets.

Hard skills

Hard skills is the toolset you’ll have to solve your problem. We can break this down into:

1. Tools
2. Business knowledge

What is the best way to know the tools you need other than checking job openings?

Soner Yıldırım analysed 2K+ job ads and he concludes that the most wanted skill is SQL, followed by Python.

Photo by Soner Yıldırım at Medium

If I would need to put into percentage the tools I use at work I would say it's 50% SQL, 25% sheets and 25% Python. However, at the same company, we have analysts with totally different distribution than mine. Some, for example, don't have Python knowledge and can still achieve good models and analysis.

For me, Python opened up the solution workspace. It taught me different ways to look at and solve the same problem.

Overall, the best Analyst is the one that applies the right toolset for a certain problem. For some, a spreadsheet works better than a complex Python model.

I've put together a list with few free courses you can pick up SQL and Python skills.

On top of those, I also believe practice makes it perfect and for that I recommend HackerRank. There you can find SQL and Python practical exercises. If you are the "on the go" type of person I love an app called "Py" it's practical and I used to do the exercises on the bus while going to work.

However, code is not all you need to know. You also need to understand the business.

I currently work as an analyst for an operational area. As I was working in operations before, a lot of the concepts come inherited to me while for other analysts, working at the same area, might have some homework to do in order to better understand the problem. Try to think about what is an area you are interested in and acquire technical knowledge on that too. You don't need to be a specialist but a general understanding is really helpful for your analysis.

Soft skills

In short, the soft skills you need to succeed on this job are:

1. Prioritisation
2. Organisation
3. Problem Structure and Understanding
4. Communication

Here the challenge begins. Most of the Data Analyst courses you find out there don't teach or even list those skills. This section of the article will be mainly based on my own experience (and some research too).

I was so excited when I first started as an Analyst that I would pick up almost all projects that came up. The result? I couldn't meet all the deadlines and ended up working extra hours to not let people down on the first promised delivery date. Here I was lacking in two things: prioritisation and organisation.

That's something I believe I am much better at today. Three things helped me with that:

1. Long term planning
2. Short term discipline
3. Better effort estimation

Now every start of the quarter I set up a project timeline. For example:

In order for me to have a full working forecast model by the end of the quarter I need to:

1. Set up our dataset
2. Analyse the data set (is there any seasonality? outliers?)
3. Clean the data set
4. Develop the model
5. Test on production
6. Write documentation

With this in mind, I know what I need to do every week. So if my goal for the week is to set up our dataset I need to allocate a certain amount of hours for this task. Once that's allocated, I have clear visibility on the numbers of hours left to perform other tasks.

Be careful to not be too optimistic. I always have blocked hours to help the team with outstanding questions, issues and for ad-hoc maintenance.

I then built a spreadsheet to schedule everything I had in the week: lunch, meetings, projects. It automatically calculates the number of available hours and this gave me the visibility to either accept or reject ongoing requests. If I have the time, I can do it. If not, a ticket is created at my backlog.

In this same sheet I calculated the effort I thought a certain task would take and later on compared that with the actual effort. With time, I started getting better at this estimation.

If you would like to try this method as well, you can find the full sheet on this link.

If you are an analyst working at a data-driven company, you might know that a lot of people reach out to you saying: "I really need the data x, y and z". At first, I used to see those requests as tasks and sometimes to get data x, y and z would take a full day or even more and this used to delay my other deliveries.

What upset me was that quite frequently this data was not being used later on, nor leading to impactful initiatives. On this scenario, I was lacking problem understanding and communication. Instead of only performing those tasks I should've asked follow up questions so they could describe their problem instead of bringing me what they think is the data solution for that.

So why soft skill is so important? Solely knowing all the programming languages will not make you solve the problems you need nor bring customer impact. From the stories I told above, knowing how to write a SQL query or a Python code was never the problem but, in the end of the day, I was still failing.

2. Work at some real life problems

To develop the soft skills things need to get real. There are several ways you can acquire those but I believe they are mainly:

1. Online competitions
2. Online courses with real life problems
3. Do analysis on your own (for the company you currently work or at some topic of your interest)

If you are the competitive type Kaggle is the platform for you. There you can compete with other participants in order to solve a problem and develop data science skills. Some of the challenges is out of the scope of the Data Analytics field but I would still recommend you opening up the data set, cleaning it up and checking the forums to see how other participants is solving similar issues.

Some courses like the Nanodegree at Business Analytics from Udacity contain real life projects for you to work on. By the end of it, you would've acquired some practical skills already. The down side of it is that usually the courses that contain those are paid.

If you would like to do things on your own and already work at a company, try to do some impactful analysis that would help the team. If there is an analyst team at your workplace that's even better, as you can ask them for guidance and offer support to start practicing the skills you are acquiring from the courses.

If you don't work, don't worry. We have a vast range of data sets available online and I recommend finding a topic you like and writing down some medium articles with analysis and recommendations based on your work.

Try to already start estimating the effort for the tasks you will need to perform here so you can combine the soft skill practice as well.

3. Apply (and get) a job

Now that's with you 😉

My last tip is for you to get into a job ad website and find some dream jobs. What are the skills they are asking? Which ones do you have? Which ones do you still need to acquire? This will be especially important for specific domain knowledge you might need.

Hopefully, these tips will help you get started and excel at the Data Analytics path.

Thanks for reading!