How you built your project:
I was our first Datathon. We have formed a team of friends of the UPC. We have analyzed the subject of the Accenture challenge to understand the different issues. Then we looked at the data to see how to process it and then use it. The data was not harmonized (strings instead of numbers, different city names depending on the files...). So we had to do a pre-processing to improve this. Then we chose a model, trained our model to finally predict on our test file if the packages are late or not.
What you learned:
For some members of the group, we had only done an introduction to maching learning. With the help of others, we were able to learn more. This challenge allowed us to progress. Moreover, working on the same project during 2 days allows to work better in team.
The challenges you faced:
We had to deal with the pressure of a time limited project. Especially on Sunday towards the end of the project, we saw how much we still had to do and how much we could have improved and it was a little frustrating.
Log in or sign up for Devpost to join the conversation.