Francisco Ribeiro Mansilha's journey in Kaggle competitions

Can you describe the project that you have been working on? 

Francisco: ‘The project I have been working on is a part of an extracurricular activity called “the Kaggle Club”. Kaggle is a platform that supports the learning of Artificial Intelligence (AI) and Machine Learning (ML) by hosting data science competitions.’  

‘This BUas Kaggle club was set up by lecturer Alican Noyan. We could either participate individually or as a team in the Kaggle competitions. There are around 530,000 active members worldwide, including both students and professionals. Each competition has a specific theme, a provided dataset, and some information about what needs to be predicted.’  

‘The task for this competition was to create a ML model that could accurately predict the desired outcome using the given dataset. The dataset was generated from a deep learning model trained on the Predicting Critical Heat Flux dataset. I submitted a ML model file to the competition, and Kaggle evaluated its correctness and accuracy based on their specified criteria. These competitions typically span two weeks, during which everyone shares ideas, tries different models, and enhances the dataset through feature engineering.’ 

What was your role in this project?  

Francisco: ‘For this particular competition, I worked individually rather than as part of a team. My role involved analysing the provided data, building models to determine the importance of different features, and trying to understand the characteristics of the dataset. I conducted research on the domain to enhance my understanding and shared my progress and ideas with other participants online. We had discussions where people shared their work, allowing everyone to learn from one another. In the future, I would like to participate as a team to experience collaborative work.’ 

Is there anything you would have done differently in this project? 

Francisco: ‘In retrospect, there is something I would have done differently. At the end of the competition, when you have to submit your work, you can see the results per file. In the end, the file I submitted wasn't the best file I had uploaded. But still, I made number 43 position out of 700 teams!’ 

‘To learn more, I will examine the submissions of the top performers to learn from their approaches. By doing so, I will gain valuable insights to improve my own models for future competitions.’ 

How does participating in Kaggle competitions benefit your career? 

Francisco: ‘Having a Kaggle profile can be interesting to build your career as it demonstrates that you possess knowledge in modelling and are actively involved in the data science community. By participating in competitions, you prove your interest in the field, your eagerness to learn, and your ability to work with data. Kaggle also provides opportunities to share your own code and datasets, further showcasing your skills and contributions.’  

What are your plans for the future? 

Francisco: ‘I had a lot of fun participating in the competition. Based on the model I build, I would be able to create an app that would ask for a few values, such as pressure, to make a very good prediction of the value in watts per square meter (W/m2). But that wasn't part of this project.’ 

‘In the coming year, I plan to start sharing my projects and build a portfolio that showcases all of my work. I am particularly interested in pursuing a career as a data scientist or an ML engineer. Additionally, I aim to participate in more Kaggle competitions since I find the competitive aspect to be highly motivating and rewarding. This recent competition was my third, and I believe that engaging in such challenges helps to refine my skills and knowledge in the field.’ 

Do you want to know more about Francisco? Find out more through his LinkedIn profile or his Kaggle profile.