Working on a real project is one of the main tenets of the Data Science Bootcamp from the very beginning.
The demo day has already been called, so now we can finish the series with final projects, through a presentation of Team 3 from the first student’s group at the Bootcamp.
Similar to Team 1, Srna, Ema, Tony, Filip and Petar feel like working on a Computer Vision Model, but with their own plan and methodology. As individuals, they all come with different experiences in the world of Data Science, but collectively they developed an advanced project that is closely related to machine learning.
If you want to work on projects like this one and add it to your portfolio tо provide you great advantages in the labour market, apply to the Data Science Bootcamp.
Brainster: Colleagues, behind you, are 12 months of hard work, several smaller projects and the final project, with which you completed the Data Science education process. Congratulations on everything you have achieved during this period. How do you feel about the newly acquired skill set and the achievements of this year?
– Thanks for the congratulations. It was difficult to maintain concentration all year. But the goal justifies the means and in the whole adventure, we gained skills that we had not even thought before that one day we will work with understanding.
Brainster: Speaking of foreknowledge, what skills did you have before entering the world of Data Science?
– We all had different skills because we came from different fields of work. What we had in common was that we were all absolute beginners in programming and writing code, while data processing in Excel was a daily routine for most of us.
Brainster: You as well as Team 1 were in charge of making a Computer Vision model that detects age, as well as whether a person wears a mask or not. In the end, both teams have a similar result, but there are differences in the methodology, organization and realization of the project. Tell us a little more about the whole process.
– We agreed that every team member should develop a model. We decided to present the best result as a collective end result. For this purpose, we made a table where we constantly monitored the progress of the models and had final control over the choice of model.
– We decided to create the dataset by collecting data from the Internet in combination with GANs. Then we used a special code with which we added masks to the generated images, so we made a selection of images and divided them into 4 categories. We used Keras and TensorFlow for the model architecture and got the best results with ResNet152.
Brainster: Colleagues from the other teams agreed that the creation of the dataset is a laborious process. Was that the case with your team?
– We honestly made the dataset easy, because we used GANs. We shared the tasks of where and how many images to enclose in the dataset. With good teamwork, things go easier. Our only difficulty was finding elderly people with masks, but we quickly solved that as well. However, our mentor insisted that we go through this part of the process to go deeper into the specifics of what a quality dataset means.
Brainster: How hard was it for you to master the technical toolkit (Python, Keras, Jupyter Notebook) in the second part of the Bootcamp? How much working on the project has helped you to be more confident in working with these tools?
– The quarantine happened just when we were starting the Python Programming module. Then we switched to online teaching. Normally, some of us had doubts about how things would go in the new format. But it turned out that the instructors were absolutely ready for the challenge. We as a team are sincerely grateful for their dedication. Their determination to unreservedly transfer their knowledge and adapt the teaching online without compromising on quality. With good instructors, every obstacle is easier to cross.
Brainster: Did something in the project development process turn out to be harder or easier than you expected?
– As mentioned, we decided all of us to be involved in all activities. This way each of us gained experience in working on all aspects of a project. Despite that, everyone had special tasks, we estimated that we should go through all the processes in the project together. We were helping each other and in the end, we chose the best suggestions as solutions to the problems.
Brainster: Which Bootcamp module unexpectedly “tired you out”, and where did you go easier than expected?
– The program of the Bootcamp is very nicely thought out, so we did not feel classic fatigue from some topic. The classes were dynamic and each module had its own specific weight. Statistics and Mathematics is one of the longest and most complex modules, but we would not say that we had any difficulties. Not only did we remember some of the things we had learned before but we also learned a lot of new things, which is amazing.
Brainster: In what domain of data working would you like to focus further?
– Most people aim at Data Science, of course, but the Bootcamp program has prepared us for work in several domains. We believe that each of us would easily find ourselves as a Data Engineer, Data Architect, ML Engineer or BI Analyst.
Brainster: From this perspective, what advice will you give to new students at the Data Science Bootcamp? Is there a recipe for successfully mastering the program and if so, what is it?
– Commitment is the key, but we all know that. The recipe for success is to stay diligent and curious, constantly testing models and learning new things. With Data Science the work never ends and the learning material is literally unlimited.
Brainster: Once again, congratulations on behalf of the entire Brainster team. What are your messages to those considering practice and a career in Data Science?
– Thanks for the kind words. To be real, if you can dedicate yourself and take your time while being ready to enter a whole new world where data is the beginning and the end, then there is nothing to doubt. Data Science is a great option for a career, and we would say that it changes the way of thinking.
If you have ever thought about a future-proof career in Data Science, have a look at our remote Bootcamps.
