Nothing hits hard like a proof of concept!
Finally, after 12 months, filled with a lot of work, lectures, assignments, and exercises, the first group from the Data Science Bootcamp completed the most intensive and complex program offered by Brainster.
At the end of the Bootcamp, the 18 participants were divided into 4 teams so they can prepare for the final projects. They have all worked on advanced projects, such as Machine Learning, Computer Vision, and Natural Language Processing. The projects are stunning for individuals that have only a one-year of experience in Data Science.
Today we will meet Team 1 that has been working on the Computer Vision project. This project is a model that recognizes people with or without masks in real-time and determines their age. The team includes a trio that came to us from NLB Bank – Ivan, Teodora, and Dimitar working in the business/financial analysis sector, Angela – Project Advisor at the Ministry of Finance, and Nikola, an active student on Faculty of Computer Science & Engineering.
Brainster: First of all, congrats on the program completion and on a successfully completed project. It may sound like a cliché question, but how does it feel to complete this 12-month, intense, and exhausting program?
– Thank you for the congrats. It’s а quite a good feeling. The 12 months passed quickly, and we achieved something that we honestly did not hope to succeed at that time. We remember the beginnings when Brainster told us to get ready for 12 months of hard work and so it was. But you realize, in the end, that it was worth it.
– We have mixed feelings about it. Happiness, pride, and privilege for the journey that took us out of our comfort zone. It made us competent to solve big data problems. To work with databases, to create predictive models with machine learning and deep neural networks that have a huge application that is still sufficiently implemented in Macedonian business practices.
On the other side, the crazy pace of a year of learning suddenly stopped. Now we have to challenge ourselves to apply and share the acquired knowledge. It was a wonderful experience, worth all the money and all the nerves!
If you want to work on projects like this one and add it to your portfolio tо provide you great advantages in the labor market, apply to the Autumn Data Science Bootcamp.
Brainster: You’ve been working on a project that personally contributed to the fight against COVID-19. You have created a Computer Vision model that detects people’s age and whether a person wears a mask or not. Tell us a little more about the motive and preparation for this project?
– We agree that everyone should personally contribute with their knowledge to something that can change it for the better. So at the beginning when we find out what project we will work on, all of us on the team agreed that we want to personally contribute to this fight that affects all of the people on the planet. Wearing a mask is considered to be an important tool to slow down and prevent the spread of COVID-19. Especially if it is universally used in all social spheres. The advice and recommendations of international authorities and many governments range from voluntary to mandatory mask-wearing in public. This means that a prolonged pandemic imposes a new way of living and learning new habits. Monitoring compliance in public is an expensive operation.
Therefore, we believe that our code as a free resource can be implemented to identify a person wearing a mask in real-time in a surveillance system. This will help to monitor the behaviour of wearing a mask in public places, such as shopping malls, supermarkets, institutions, etc. We strongly believe that each individual should strive to contribute to suppress the spread of Covid-19 and help to end the pandemic. So, this is one of our ways to contribute 🙂
Brainster: The project was planned so you can create a dataset by yourself, which we believe was a serious challenge. How did the creation process go, and with what dataset did you end up?
– To create a dataset was one of the most difficult tasks and the biggest challenge we have. In communication with our mentor, we had two options on how to create the dataset. We could choose an existing set or create a new, unique one, so we decided on the second option – to collect images that we had to take, without using ready data and datasets that were already created for this purpose. We did this so that our database contained 80% of the actual images we took as a team. Adding a face mask in a synthetic way (with code) was not used as a method, so we created a real and authentic database. We are very proud of the team effort.
– The process of collecting photos sometimes included requests to our friends to take photos and submit photos to us on social media. However, it was also a fun process, when people didn’t know exactly how to take photos or submitted funny photos…
Brainster: Besides this, what other challenges have you faced along the way?
-Well, model training was certainly a challenge. For that purpose, all team members were in charge of training on, as many algorithms as possible, so we can get a model with the highest precision.
-The selection of an appropriate neural network, the complexity of the layers and the adjustment of the hyperparameters versus the model training speed, the resources required by the network, the stability of the model, and the accuracy/precision/errors as measures for evaluating the model were the requirements that had to be met and brought to the optimum and best mix.
Admissions are open for the Fall Data Science Brainster Bootcamp. Join us and future-proof your career by building a top-notch and job-ready portfolio.
Brainster: This is your first Data Science project and the first point that awaits you further, by working in a production environment. For extra pressure, there was a time frame of just 3 weeks. How important was communication within the team for the success of the project? How did you manage to sync?
– As for the first project and in general, working on something completely real and applicable to the production. we are very satisfied with how the whole process went. This has only further motivated us to continue what we want to do in the future and that is Data Science.
– We have great communication. We, as team members, had excellent communication and cooperation before this project, so we decided to cooperate for the final project. After the first week, we simply anticipated each other’s behaviours, which was very important for mutual cohesion, which would complement the already high productivity for the entire duration of the project. It is a pleasure to work in a team, in which you know what to expect from each member.
Brainster: For the needs of the project, you used Python, Jupyter Notebook, Keras, TensorFlow, NumPy, OpenCV, etc. How does it feel to master this toolkit based on your skill level before enrolling in the Bootcamp?
– What seemed impossible to us 12 months ago, has now become possible. Before we started with the Bootcamp, it seemed that programming in Python and mastering all libraries would be a big bite for us. But with the whole Bootcampconcept and the capacity of the lecturers, each module is designed to gradually introduce you to the world of programming. And indeed, when the Python module came along, it no longer seemed impossible, as it had before.
– Of course, an analytical mind, prior knowledge of mathematics and logic, consistency, and desire for lifelong learning. These are the prerequisites that should be met by anyone who wants to master such a strenuous program.
Brainster: Brainster is already discussing a plan to implement your model, but where do you think it might be most effective?
– As we mentioned before, there are many possibilities for this solution to find applications. Like, in all public places, in front of bank entrances, closed shopping malls, institutions, large companies, ZOOs, amusement parks, cultural facilities (theatres, MOB, philharmonic, MKC, etc.). All those who have cameras for Supervision can use our solution to perform real-time surveillance which will regulate the mask-wearing. It will directly influence and contribute to the protection of public health from the spread of COVID-19. Moreover, the very use of such a solution in as many places as possible for a short period. This can raise awareness among people and change, for the better, the habits of not wearing a mask.
Let’s go back to the program. From experience so far, Statistics and Machine Learning along with Python have proven to be the most challenging modules.
Brainster: What would you recommend to current students to make it easier to cope with the most difficult part of the program?
– Follow the program regularly, timely, and quality completion of all tasks assigned by the mentor. And for Machine Learning – exercises, exercises, and only exercises! You can find on the Internet, models for many datasets, with all kinds of algorithms. They just need to be found, implement, and build (to try) different architectures to come up with the most efficient solution.
– Efficient time management, a lot of reading, and work.
Apply for the Data Science Bootcamp starting this Autumn. Secure your spot!
Brainster: Some of you continue to work as part of the Data Science hub. This time you will work on a real project for the Macedonian Stock Exchange. How much has this practical work approach helped you to master the key Data Science skills?
– The Bootcamp gave us a great foundation. During the education, we worked on many projects, with real datasets, on real and current problems. This project will be a natural continuation of our educational process to strengthen our skills and build more experience.
Brainster: Finally, on behalf of the entire Brainster team, we would like to thank you for your participation at the Bootcamp and for the project you have left behind. For the first time, from the perspective of someone who has taken the difficult path, what would you say to anyone considering enrolling in the Bootcamp and pursuing a career in Data Science?
– Currently, Data Scientists are among the top 5 most wanted professions in the world. Do you need any additional comments? 🙂
– Well, the Bootcamp is not naive at all. It requires maximum effort and dedication. But if you want to do it, nothing is impossible. Indeed, from this point, in discussion with others, we came to the same conclusion: We all enjoyed what we have been doing for 12 months. We don’t even regret some sleepless nights, working on a project, because we know it was all worth it!
* Detailed information and overview of the project that Team 1 worked on, you can find at the following link:
If you have ever thought about a future-proof career in Data Science, have a look at our remote Bootcamps.