At the beginning we had a hard time choosing the challenge that motivated us the most. We were looking for something related to AI and data science, so we finally decided on the BUNGE challenge.
Our project is not a finished and fully functional programme. It is presented more as a start of a programme that can answer questions related to the database provided by BUNGE.
Due to our low level of AI, we decided to implement three phases for the project. The first phase focuses on transforming the question into interpretable parameters for phase 2. The second phase focuses on different methods that are used to trim and simplify the database in order to be able to deal with the databse for the answers (the output), i.e. the third phase. The third phase is in charge of producing the answers to the questions, and is the least developed part of the project.
The most notable challenges we ran into were: how to treat the dataset, how to interpret it, and how to interpret the input in human language. In other words, we encountered challenges in all phases of the project.
The accomplishments that we're proud of have been mostly related to understanding the proposed objective and transforming it into code. It has been a challenge to find a way to solve the problem with the tools we know. And, although it may not be the most optimal solution, we're proud to have come up with a workable solution on our own.
We have improved our learning of Python in general, as well as learning new libraries such as Pandas, re, etc. In addition, we have seen the importance of efficiency in human facing programs.
Finally, we do not discard the idea of continuing with this project and applying, in the future, the methods learnt in our studies in order to obtain better and more optimal solutions.
Log in or sign up for Devpost to join the conversation.