Rossmann Store Sales Project Report

(from Kaggle Competition)

Project description

Rossmann is a chain stores in Europe. And this project is from Kaggle competition Rossmann Store Sales. I need to predict Rossmann stores' sales according provided data, like the promotion, competition, holiday and etc.

Requirements

In this project, I use Python for data explore, visulization and featuring engineering. Python version is windows X64 3.5.5. And involved packages inlcude:

• Jupyter Notebook
• numpy
• pandas
• seaborn
• matplotlib
• xgboost

Run

1. To run this project, you should include sub directory: input. The Input directory has data downloaded from Kaggle and unzip to local: train.csv, test.csv.
2. In Jupyter Notebook to load “Rossmann Store project for report Eng.ipynb” and execute it. The executation time needs about 7 hours。
3. The result is "xgboost_rossman_submission.csv" and can be uploaded to Kaggle.

Result

The result of local test set: RMSPE: 0.109707

Kaggle Private Score 0.11518

Kaggle Public Score 0.11636