3 Unexpected Work Activities as a Data Scientist

 


Many people think that the only work data scientists do is coding and develop machine learning models. Coding skills maybe are the reason we are employed, but the reality is different. We have many unexpected works to do outside of coding, and only known after becoming data scientists.

That is why in this article, I want to explain my experience as a data scientist that our work specification is way bigger than many people expected. Let’s get into it!

1. Convincing People

What is this supposed to mean? We data scientists need to convincing other people? What are we convincing for, and why is it necessary?. This kind of question might present in your head when I told you that data scientists need to convince people — but convincing people is our work.

Data science, Artificial Intelligence, and Machine Learning is a newcomer in the industry. In the past, we already got used to making a decision based on instinct and whatever made sense at the time. With new technology present, we can process all the available data to make a data-driven decision and not rely solely on instinct. However, just like any new things presented, many are skeptical of the data science prospect.

Explaining data science prospects is a Data Scientist job — we need to convince people that our work would benefit their work, benefit the business, and benefit the company. Why do they want to use your insight and machine learning model if their old way has worked? This question is a common hurdle in our everyday work.

In my experience, I have met with a lot of resistance from business people. People who did not understand what we did, people who did not want to change their ways, and people scared of new things. I can’t blame them for not understanding, as machine learning models were akin to magic for non-technical people. But, convincing our work is something that data scientists need to do — if people don’t want to use your insight or model, it is a useless effort.

Am I always successful in convincing people to use data products or machine learning models in their work? The answer is No. There are many times that people deny the usefulness of data science. How is the ideal way to convince them?

There is no golden way to convince people because every company has its own culture, every business has its way, and every person has its personality. One piece of advice thou that I could give is:

Try to convince people by presenting the usefulness of your model using the data with the aim to make the other party curious. Don’t aim for instant conviction on the first meeting, but try to convince them for another meeting with a “study” of their business problems.

You can’t convince people without adequate proof; that is why you need something to support your claim. Data scientists would always find people who are inconvincible, but you should always appreciate more open-minded people.

2. Leading

At first, data scientists might only work to develop the machine learning model to solve business problems, but with experience came more responsibility. With time, this responsibility would evolve where you would need to lead many things within your work.

Take my personal experience as an example. The company trusts me to develop a machine learning model to help solve the business problem in my early day. For a beginner, I would not understand much about the business and the data. However, after sometimes I would know what is required for my machine learning model to help the company — it needs a combination of many departments: sales, marketing, data engineer, IT, digital, and more. Because I am the one who knows the best how to make this project succeed, I am naturally becoming the project leader and need to maintain all the variables for success.

In our data science work, we would need to lead at some point, even if we are in the Junior position. The scope might be different for each person, but you would experience leading some way or another. You might need to lead the project and connect with many people, leading your team or even leading the direction of the data success in your company.

Many data scientists don’t like to do this as they think it is not their job description. However, if you want to advance your career and getting somewhere with your model, you need to learn leading. Additionally, the insight from your analysis and your model development is valuable — and you are the one who knows best how it would impact the business. If we are not the ones who lead this change, who else would be?

Few tips from me if you want to have more experience in leading as a data scientist:

  1. Take the initiative. Don’t just expect things would work out for you, but make your own path,
  2. Don’t shy away from the responsibility given to you. Work given to you is a sign of trust and a chance to show your leading ability,
  3. Communicate your wishes to your boss. If you want the experience, you should discuss it,
  4. Active in the activity outside of your work. You might work in the data department, but there are many people within your company. Try to get more people outside of your usual working circle; it would help.

3. Business Translator

Data science is a new thing in the industry, and many people still don’t understand how useful it is for the business. This sentiment means data scientists need to become the middle people between the data science concepts and the business people.

I am always excited when business users come with their problems and need our help. Why am I excited? Because I could be creative to help their business. In my personal experience, business people sometimes have a great idea to solve their business problems and somewhat know what to do. Still, they are not sure how to translate it into the technical execution. This is where we, as data scientists, come to help them analyze what they want and break it down into a viable solution using our creativity.

For example, the company might need the analytic team's help to produce better sales leads because their current process hasn’t managed to reach the target. We could help them by analyzing the root cause and develop a machine learning model to produce the leads. Sounds easy for data scientists, right?

However, imagine yourself as the business people presented by data scientists by this kind of approach:

“From the analytic team, we would analyze the data to see if there is a strong correlation within the features and we would develop an ensemble machine learning model with precision target 80%.”

Data scientists understand the terms above, but non-technical people would not understand. They do not know the technical jargons we are uttering, which only makes them more confused. The consequences of confusion usually come to two things:

  1. People avoid using the analytic approach because they did not understand,
  2. People accept the approach, but the execution often deviates from the course because the concept has not been established firmly at the beginning.

For the business people, we might transform the technical terms above into this:

“From the analytic team, we would analyze the data to try finding out the reasons why the sales are down and we would develop tools to help predict leads with high chances to buy the product. Ideally, the leads would contribute to 10% of the revenue target.”

The above sentences using less technical terms and aligning the target more with the business target.

We, as data scientists, need to translate the technical solution to the business people in a way that people would understand. It takes practice to fully translate the data science requirements to the business language but not impossible. There is advice I could give to help to learn the translation skill:

  1. Avoid technical jargon as much as possible and use the business language,
  2. Don’t overpromise. Data scientists are capable of developing the model, but many variables affecting the success, so our promise could be interpreted as something else,
  3. Dig as much information from the business people. Align your thought with them and ask questions if you don’t understand.

Conclusion

Data scientist is employed because of their analytical and coding skills. However, their work activity is more than that. In this article, I have explained three unexpected work activities as a data scientist. They are:

  1. Convincing People
  2. Leading
  3. Business Translator

Post a Comment

Previous Post Next Post