break into data science

Data Crunch: How To Break Into Data Science

Aug 15, 2018

How do you get your first job in Data Science? If you do a search on it, you’ll find lots of answers.

Some would advise you to start as a data analyst, a data engineer or to go into business intelligence first… but there never seems to be a real answer.

There has been a number of data scientists in Malaysia, trained by Dr. Lau Cher Han, who are now working as data analysts & scientists in prestigious companies such as AirAsia, Fave and more. A few of the highlights in this episode:

  • If you need a technical background to become a data scientist.
  • The different types of data scientists. (Type A or Type B)
  • Which data science tools should you learn?
  • How do companies in Malaysia hire a data scientist?

Welcome to Data Crunch

Reuben: Welcome to Data Crunch, where we talk about big data, data science, data analytics, machine learning and AI. In our first debut of Data Crunch, we are going to talk about breaking into data science, a career in data science. Dr. Lau is a chief data scientist and founder of LEAD.

Can you become a data scientist without a technical background?

Dr. Lau Cher Han: I get asked this question a lot. Actually, you don’t need any technical or programming background when it comes to data science. In data science, we categorize people as – A type data scientist or B type data scientist.

A type data scientist would be somebody like yourself, Reuben. You are analyzing numbers and it’s your day job to look at reports and statistics.

B type data scientist would be more of a ‘builder’ type – so someone with a programming background, who moved into the career path of data science. So to be honest, the field of data science is vast. It doesn’t matter if you have programming knowledge or not – everybody can learn data science and eventually become a data scientist.

How do I know if I’m a Type A or Type B data scientist?

Reuben: How do I find out if I am a type A data scientist person or type B?

Dr. Lau Cher Han: Basically, if you have a programming background and you know how to write programs, then you’re most probably type B person – you’re a builder. For example, people who build programs or build models. These are people who enjoy fixing problems and tweaking parameters.

A type A person would be somebody who is already very well versed in analytical software such as SPSS or Excel. But they are not used to using scripts to automate a task. Or they are less into things like databases or ETL processes. That’s a simple way to determine the type of data scientist you are.

What’s a good pathway to go into data science?

Reuben: If I’m someone without any formal education and no experience what’s the best place or path to get started with data science?

Dr. Lau Cher Han: A lot of people who are interested to break into data science, will have this question, “do I need a degree?” But data science is almost a new field. Before that, it was just a concept. And within data science, you have a lot of different things like machine learning, algebra, mathematics, a bit of programming and a bit of data science, such as database skills for example.

So when it comes to getting a data science degree, most of the degrees provided by universities are generally new. You don’t really need to have a data science degree but ideally, if you have a relevant degree, that will be good enough.

Besides, there are many programs out there, such as boot camps or certifications that you can attend. They’ll help you to get onto the right track to begin your data science career path.

How much is the salary of a data scientist?

Reuben: I’m really curious. How much is the usual salary of a data scientist in Malaysia or the world? What kind of salary can I expect as a junior data scientist, intermediate data scientist, or maybe a chief level data scientist? Is it a good career switch?

Dr. Lau Cher Han: You must have read the article, where it says data scientist is the sexiest job of the 21st century, right? Let’s start at the top. Whenever people talk about ‘chief’, it usually means the CEO, CTO level. Whenever you’re at a chief level you are usually less involved in the dirty work, but more involved in management.

So you’ll be working with a team and managing a group of associate data scientists or junior data scientists. In my case, in I-Stream, and in my own B2B companies – we have 5 to 6 data scientists for one project. As a chief data scientist, I have to manage them, and different KPIs and stuff to look at.

Usually, at the chief level, the salary you’ll be looking at would be around RM100,000 – RM150,000. If you are in Australia or the US most of the time, then that would be in its local currency.

The entry-level data scientist or data analyst, usually makes a good RM4000 – RM6000 on average. My students who work in corporate companies earn an average salary of RM6,000 to RM8,000.

Will companies ask for certificates when hiring data scientists?

Reuben: When it comes to data science and getting employed, won’t companies ask for a certificate? If I don’t have a data science certificate, what then?

Dr. Lau Cher Han: I’ll usually make sure that my students make their own portfolio. In fact, the HR person from most companies usually doesn’t know what sort of questions to ask, when hiring a data scientist. So they then ask for a certificate because its more like a ‘minimum requirement’ to enter the company.

Take me as an example, my bachelor’s degree is in IT and I picked a major of database, followed by my masters and PhD, where I focused more on AI and machine learning.

More importantly is whenever you attend interviews, you have to show the company the values you can bring to the company and if you are able to do the job. And back to your question, most of the time my students apply to work at startup companies, they found that the degree itself is not really that important. I wouldn’t say totally not important, but less important. Skills and know-how usually matter more.

Which individual characteristics will make a good data scientist?

Dr. Lau Cher Han: There are three criteria here.

1 – Good in Communication

As data scientists, you usually need to communicate with a lot of people. I’m not talking about normal communication or interpersonal skills, but the type of communication you have with clients to understand their pain points, to find out where the data is, what are the things they are looking for and how to solve their business problems.

And at the same time, you have to be able to do good presentations and tell compelling stories to stakeholders and to your audience. So, the number one key characteristic of a good data scientist is good communication skills.

2 – Open-Mindedness

In data-driven companies, decisions are no longer made, based on guts. Traditionally, experienced executives from companies believe they can use their guts to make decisions. But today, most decisions have to be made using a data-driven approach using more scientific methods.

As a data scientist, be open-minded to look at data and think. What does this data mean? What is it telling us? And from here, what are some experiments we can do? What is the hypothesis? Generally, any type of feedback is good feedback

3 – Curiosity

You must be curious like a baby to be good in this field. You can’t just look at the graph and make conclusions. Oh this the outlier, this is the reason, etc.

Whenever there is an outlier in your chart for example. Say, you plotted a chart for an E-commerce company and see that a customer usually spends about RM200 to RM300 in every credit card transaction. Then one day, he suddenly spends two to three thousand. Is it a fraud transaction or is it something else?

Be very curious about different things that come out. Never jump to conclusions, but always be looking at the broader picture and be very curious whenever something exceptional comes up.

What data science tools should I learn?

Dr. Lau Cher Han: There is no one right tool to use. But I would say, Python is the more suitable and popular tool that most data scientist would be using.

Python and R are both open source and free, so you can just download, use it and implement all sort of models with it. There are lots of scientific packages that are already implemented, frameworks that are already there. So yeah it’s very easy to get on track.

Any tips to take a data science interview?

Reuben: So, let’s say I’m going for a data science interview tomorrow, what should I prepare? What would the interviewer ask me?

Dr Lau Cher Han: I think I covered some of it in the previous question, whenever you go for an interview, the HR or the interviewer are usually not so clear about the job scope of the data scientist. Most of the time, we data scientists, know more about data science, and the person that is going to hire you to know more about their company.

My advice is would be to try to take a non-conventional approach rather than just sitting down and asking them about the salary packages. I would prefer you to take the pro-active approach, to ask more questions. Ask them about their business pain points, the business questions that they are facing and how you can help them to save cost or to increase revenue. From there, you show that by using the skill sets that you have, you can bring more value to the company. This is a much better approach rather than the traditional interview style that most do.

Which company should a junior data scientist look for?

Reuben: What are some companies, in your opinion, suitable for a junior data scientist to join?

Dr. Lau Cher Han: If you’re a junior data scientist, I’d suggest you to join a startup. At least not a too new startup but a well-funded startup. This is because they will have lots of data for you to play with, but at the same time, they don’t know what to do with all the data.

If you’re someone who is already in the mid of your career, you should already know what you want to do. I’d assume you’re looking for job stability and is comfortable within a corporate environment – so look into different industries like banking, insurance, transport and also telecommunication companies. These are the industries that are heavily data-driven. By joining them, you ensure that you always things to do with data.

Do note that many bigger traditional corporate companies have data, but they are scattered around. So, when you go in, expect to do less of a data scientist job but actually spend 80% of your time doing data gathering, crawling, and cleaning.

What are other data science questions you have that we did not answer? Let us know in the comments section below. Like us on Facebook and subscribe to our YouTube channel

Brought to you by

The LEAD team believes in growth. The question we ask every day: How can we help our students achieve more?

Get the scoop on the latest stuff.


Ultimate Guide to Data Science

Recent Posts


Submit a Comment

Your email address will not be published. Required fields are marked *