From Being A Sales Rep To Being A Data Scientist

25Sep - by aiuniverse - 0 - In Data Science

Source: towardsdatascience.com

The cold calling terrors

There are no words for how much I hated cold calling. I did not want to do it and the people on the other side of the phone did not want me to do it either. It was failure and rejection at the other end of the line 95% of the time. I was not the naturally super friendly and chatty type on the phone, so each call took a lot of efforts from me. But someone had to do it and I had to keep trying.

Being in Business Development did not only include cold calling, most of it was account & relationship management. As I was selling online marketing campaigns, a small part of my job involved analytics & statistics about the campaign performance. Surprisingly, I really liked presenting and explaining the data to my clients. I loved the technical part of my job, even though it was really not that significant and relied mainly on Google Analytics.

You can’t do it, he said

After realizing that Sales was not for me, I entered a Pre-Master in the Netherlands. I did not really know yet what specific Master track I was going to choose if I passed it. That Pre-Master was quite a general preparation program which could lead to several tracks about Online Marketing, Data Journalism or strangely, the new Data Science track. Since I did not know much about the topic, I had not even considered the Data Science track, yet. Someday, one of my fellow students mentioned during lunch that he was going for that track. He mentionned how difficult the Master would be for me, how likely I would be to fail that track because I had no experience in Python or in programming. Honestly, I hate people telling me what I can and cannot do. It often sounds to me like a golden invitation to prove them wrong. So a few weeks later, after researching the topic in details (and actually enjoying my statistics course), I took a decision.

Data Science, here I come“.

Learning programming as a complete beginner

Learning programming from scratch was not a walk in the park. It required countless hours spent on basic programming exercises, facing the red error messages & wondering if learning dark magic would have been easier than learning Python. I also started understanding the hate-love relationship that programmers sometimes express towards a language.

Looking back, I realized the deep truth behind what my Python programming teacher had said: “If you do not practice enough, you will fail.”. Yes, it seems rather logical, any new skill requires practice. What he really meant is that programming is not a subject that you can pass from reading a few lines of codes last minute before the exam. You need to teach yourself to think like a programmer, which requires hours of practice on different types of exercises.

Only 30% of my class made the first Python exam and with my countless hours of practice, I, fortunately, ended up in there as well. In programming, the answer always seems easy when you read it. It is another story when you have to design and write the solution entirely yourself. In short, keep practising and do not just copy-paste answers from StackOverflow, as tempting as it can be.

Machine Learning sounded so sexy

I remembering loving these two words “Machine Learning” from the first time I heard the definition. I felt so much excitement about this topic and could not wait to learn more. I thought I could impress my family by telling them that I was teaching computers how to learn stuff, even though to this day, they still understand nothing about my job.

At first, the Machine Learning concepts were not difficult. However, the statistics behind certain models had a hard time sticking in my mind. Applying formulas and understanding the logic behind them had never been my strong point. Why did I end up in Data Science again? To get a full understanding of the inner workings of certain models, I often spent additional time researching how algorithms worked. I did not always need to know the details, I just wanted to. The more I researched, the more I realised how little I knew and how rich the topic was.

Visualizations also played a big role when I was getting stuck with a certain model or concept. For instance, I strongly preferred using graphs to learn the details of the gradient descent optimization or the differences between neural networks architectures.

Using my background as a strength

A background in Business Development can seem weak because it lacks technical skills and has very little connections to the two biggest subjects related to Data Science: programming and statistics. Anyone with a Computer Science degree or an Econometrics degree will obviously have some sort of advantage when it comes to efficient programming or the statistical aspects of Machine Learning models, sure.

Nevertheless, many seem to forget that soft skills in Data Scientists are extremely valuables. Next to applying powerful Machine Learning models, Data Scientists have to tell the story that the data is hiding, and no one wants to listen to a boring, platonic story.

Surprisingly, I realized during my Master thesis that I was quite decent at presenting, and that I could easily turn my models into simple concepts that non Data Scientists could understand. I would make up grocery stories to explain recommendation systems and potential scenarios to explain the measurements that I had obtained. As a Sales Specialist, I had spent 2 years presenting products to people, telling them stories to put the products into context, limiting it to the content that mattered to them, using words that made sense to them. Explaining the results of a Data Science project was not much different.

Moreover, the fact of having experienced the commercial side of any product or project can also help greatly when it comes to bridging the gap between business and technology needs. It can be useful to have some sort of translator to prevent confusion between people who speak the same language, just slightly differently.

Knowing my limits

When making a switch as big as this one, one has to be ready to put in a lot of efforts and to look at their own limits. I am very mindful about my own weaknesses and anyone should be, even geniuses who have been programming Neural Networks since before anyone knew what it really was. I am not the best at statistics and I am not the best at programming beautiful efficient code. However, I know how to take an opportunity to learn and who to learn from, when I see someone with better coding skills or that seems incredibly comfortable in statistics. No matter how much experience a Data Scientist have, this field is in constant motion and no Data Scientist should ever be done learning.

The beauty of Data Science is also the beauty of the different backgrounds that make up the field. Data has so much to tell, and data can be analysed and worked on from so many different angles. Even though certain backgrounds will make your progress a lot easier than others, any background will bring something new to the field. It can even become a real advantage depending on the type of data you might work with. The success of a career switch is often only determined by how much effort one puts in it.

Facebook Comments