文件列表:

Hi!ckathon4 - Completion Certificate.pdf
Hickathon4_student_guide.pdf
Model.ipynb
Preprocessing.ipynb
Variable description - Hi!ckathon #4.pdf
hickathon_report.pdf

# Hi Paris Hackathon 2023 4th edition of the competitive data challenge organized by Hi! Paris, the interdisciplinary research and teaching Center for Data Analytics and Artificial Intelligence, on AI and Supply Chain ## Dataset: Over 2 million lines, each containing 4 months of sales of a defined product (anonymized), as well as extensive information including country of production, sales, target customer, macro-economic and environmental data. Note that since the dataset is large, I have not provided the data on the repository. ## Goals: - Implement a Machine Learning model to predict sales volume for month 4. You must avoid overproduction and propose measures to reduce the overall carbon footprint. - Write a scientific paper which propose a solution to optimize sales prediction while taking into account the full CO2 emissions of the value chain. - Design an app that: 1. Leverages your ML model 2. Present your result in sales prediction 3. Provides insights to improve sustainability ## Experience: This was my very first hackathon and while it being very challenging, I learned a lot from my 5 teammates and coaches (from data scientists to business experts) on how to tackle a data project in less than 48h. We split the work between implementing the machine learning model (optimizing XGBoost) and designing the app (using Django Soft UI Dashboard). ## Acknowledgments: Thank you to the researchers at the Hi! Paris research center as well as the corporate donors from l'Oreal, Capgemini, TotalEnergies, Kering, Rexel, Vinci and Schneider Electric for this opportunity to learn and gain valuable experience in the data science world.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：