Data Scientist | Three different views on a professional’s day to day life
Data Scientist is a new market profession that is on the rise and has created a new era of information we are experiencing. This article views this promising career from three professionals from the sector, who share their experiences.
Learning is continuous | Edgar Bermúdez
Being a relatively new profession, data science is hard to describe and there are various opinions on what a professional in this market should know and master. In practice, the Data Science team is made up of people with different backgrounds and abilities. My team is mainly composed of people from Mathematics, Statistics and Engineering, all with complementary knowledge and who are constantly collaborating.
The team is constantly learning. We work with many projects that require skills that need to be learned with online training without forgetting the deadline for project delivery. There is a lot of learning and software to master and it is vital to prioritize learning the software that each project requires. Learning is continuous and mandatory.
My career began with integration into an ongoing project that I was responsible for shortly thereafter. The first few days were divided between understanding the problem to be solved, monitoring project development and learning Python, a new programming language for me. After understanding the project, more than 80% of the work consisted of packing the data received from the client, understanding more than 20 different databases, treated and unified in order to extract the necessary information. Later, I went through a challenging phase in which the discoveries made using classification and clustering techniques in addition to the information that had to be presented to the client in an easy-to-understand way, the complicated art of making the difficulty seem easy.
My second project also required a lot of learning, since it required preparing research results for an international scenario applied to a national scenario. Again, part of the work consisted of only data storage in order to work with them. Here, in addition to learning new Python tools, it was necessary to learn R classification and prioritization of characteristics to reach the expected result on time.
The third and current project is aimed at automatic document identification, a new challenge that requires Deep Learning and Machine Learning knowledge as well as new ways of storing information, presenting it and working with different technologies. In the end, working with Data Science is an endless learning process.
It is an intellectual diversity sector | Luiz Nogueira
The complexity, dynamism and constant search for information are significant in the consulting environment, and requires data scientists to keep abreast of innovations for delivering efficient solutions to the customer. Getting out of the comfort zone and adjusting to the speed of learning and results delivery is how I would best describe my early career as a data scientist.
I graduated in Mechanical Engineering where I had little contact with programming. At everis, I had the opportunity to learn new languages, knowledge and over time, I was able to deepen in Python knowledge that helped me learn advanced technologies in the Data Science sector such as NLP, Deep Learning and Machine Learning. I participated in three projects where I managed to demonstrate the sector’s complexity and heterogeneity, which reflects on a unique learning curve for my career.
My team of Data Science is marked by the intellectual diversity components, making each day full of new challenges and learning. Teamwork enables the exchange of information and its alignment reflects an innovative and well-structured project. The teamwork also contributes to a light working environment, with the home office option, which allow employees to adapt their needs to the obligations efficiently.
Globalization allows rapid circulation of information and should be used with caution, always analyzing the truth of the facts. A data scientist must collect, analyze, and interpret a large volume of data to create or generate solutions to a given problem. Therefore, to keep me up-to-date, I perform online courses on safe platforms indicated by the company.
The main ways to keep up to date are are reading articles, following great researchers and companies on social networks, always keeping an eye out for new posts about new technologies. However, what is difficult about keeping yourself informed about sector innovations is undoubtedly to combine work, rest and time to read and study.
You have to like working with algorithms, methods, models and numbers: Laryssa Kato
Several different definitions of data scientist assignments are on the internet. However, with my nine months of experience, I know that the minimum requirement is to understand basic statistics, mathematics, and above all, to be familiar with analytical geometry. I know this because it was the only knowledge that I had, besides the knowledge of the Fortran programming language.
Before I was a data scientist, I was from the academic field. I have a Physics degree, where the pace and dynamics of work are very different; it took me some time to adapt to new ways of thinking. Basically, I still have a problem to look for a solution, what has changed is that now, it is necessary to find solutions fast — time is short, and we have to meet deadlines for project delivery, solve problems efficiently and meet needs.
What I have learned throughout my experience in the new profession is Machine Learning methods that are basic and necessary to understand how the structure works to create a suitable model, where everything starts in the exploration and structuring of the data. In Machine Learning, I have the most experience in the neural networks sector, more specifically, Long Short Term Memory (LSTM), which I apply to the project I’ve been doing since I started my career. I learned the exploration of data and its importance, since we can extract very important information that determines which model we should choose, how we should structure the data, and see the real needs that the project has. Most of the time we organize, structure, and clean up this data, and we do this using Python, a programming language needed in project applications that I am currently learning.
My closest contact is with the project team I work with, holding meetings, discussing problems, collecting data, updating the status on how the model is built, and other activities. The project I am working on is not just data science, we are one of the several puzzle pieces that complete a project. I do the implementation of machine learning through a neural network model. I do not act alone, I work with one other person, which I help and we learn a lot together.
What I have felt since I started this career is there are many algorithms, methods, models and problems that require data scientists to have different knowledge, since we never have the same problem to solve. Each project requires a study to help solve the problem. Also, only Machine Learning knowledge is not enough. I needed to learn SQL language to be able to search and download data from a database (hive) and know some Linux, which I had basic knowledge, due to the programs I performed when I was in the academic sector. I am now diving deep in this knowledge.
Where I work, we have a team of data scientist only and each one works on a different project and often works as I do, as part of the project, with several people. In my case, I do not have direct contact with the client, but other professionals are constantly in contact with them.
We have weekly meetings that help know the projects that are all inserted and expands the basic knowledge about things to be able to search more on the outside. What I like most is we always learn a lot, since we have courses to help with training and people who are committed to sharing their knowledge and who are thinking about the overall team win. Online courses help with the construction of knowledge, we learn about theories and put what we know into practice, which makes knowledge more complete.
In all the reports, the work routine of this profession is never the same, there is always something new to learn and problems to be solved that require more diverse knowledge in the sectors of technology and information.
Through data, it is possible to obtain information that is not evident and becomes knowledge that benefits society, optimizing processes, streamlining procedures and in general, improving the quality of life of people and providing the satisfaction of helping others.