Arvato-Customer-Segmentation

Customer Segmentation repository for Arvato Financial Services.

Requirements

Package requirements are described in requirements.txt. Additional requirement is the access to AWS cloud for Sage Maker used for hyperparameter tuning.

Project structure

Project proposal

Project proposal could be found in Proposal/proposal.pdf.

Project report

Project report could be found in Report/ArvatoReport.pdf.

Analysis

Analysis is split into multiple folders corresponding to various steps, also described in the project report.

1. Folder EDA: Contains the Exploratory Data Analysis Steps along with concating and imputation steps of data processing. Notebooks names are starting with an intiger indicating the order in which they should be executed/read.

   • 01_EDA_AZDIAS.ipynb
   • 02_EDA_CUSTOMERS.ipynb
   • 03_Concatenate_and_Impute_Azdias_Customers.ipynb
   • 04_MAILOUT.ipynb

2. Folder Unsupervised Contains the major part of this project, namely the customer segmentation projects that identifies and characterizes the market segments. Explores its meaning, and charactierizes its importance.

   • Unsupervised_Customer_Segmentation.ipynb

3. Folder Supervised Contains two notebooks for two types of models tested for prediction of successful response of a customer after being exposed to a marketing campaign.

   • 01_LinearLerner.ipynb: a baseline linear model
   • 02_XGBoost.ipynb: tree-boosted model with hyperparameter tuning

4. Folder submissions contains the final .csv files for the submissions to the Kaggle competition.