Table of Contents

1. Description
2. Dependencies
3. Installation
4. File Descriptions
5. Results
6. Licensing, Authors, and Acknowledgements

Description

This project in collaboration with Arvato a mail-order sales Company in German , is part of the fulfillment of the Udacity DataScience NanoDegree. The datasets here are provided by Arvato. But as part of the terms and conditions I'm not allowed to share and or include them in this repository. The Files included the following

- Demographic information for the general population
- Demographic information for existing Costumers
- Demograghic Attribute Informtion file 
- Demographic Attribute value information file

** The Main Goal of this project is to study both the demographic data for the genral population and that of companies customers and use both unsurpervised and supervised learning algorithms algorithms to determin the individuals of the general population that can be come potential customers. **

A project containg a jupiter notebook and related files, this is the main part of this project with has all the interaction with the datasets it 3 parts

Data Cleaning

Steps in this section include - Gather and Explore the datasets getting all important information from the dataset. - Perform Data Wrangling to generate a cleand and standardized dataset.

UnSupervided Learning

Here we use unsupervided learning algorithms to group invididuals in the general population and customers to perform our analysis eventually. Steps her involve - Reduce dimension of the dataset for easy of visualization - Group individuals into groups(cluster) based of similarity for anlysis - Draw a conclusion based on analysis

Supervised Learning

Here we user Machine Learning to perform predication on the potential customers from the general population. Steps here include - Based on the given dataset train a machine learnin algorithm to predict the potential response of individuals. - Optimize the Algorhms to get best algorithm - Draw a conclusion based on the results

---

Dependencies

- Python 3.x.x+
- Machine Learning & ELT: Pandas, Numpy, Sciki-Learn
- Model Persistence: Pickle

Installation

Clone Reoository and run in an environment that supports jupiter notebook.

File Descriptions

There are 3 main parts

• Arvato Project Workbook.ipynb The Main jupiter notebook file.
• Arvato Project Workbook.html Html Represention of Arvato Project Workbook.ipynb notebook
• The Main Data flies could not be share which part of the terms and conditions with Arvato

Results

1. See medium post here

2. After runing the Machin Leaning Alorithms an f1-score .94 was obtained, 94%

3. A web interface to test the model, inferfaces below

Licensing, Authors, Acknowledgements

Credits to Arvato for the data.

Author

Fuh Austin

udacity-datascience-capston-arvato-project