RECENT Projects
customer segmentation and prediction
The project leverages demographics datasets provided by Arvato Financial Solutions. We provide a solution to the customer segmentation/prediction problem of a mail-order sales company in Germany:
Unsupervised algorithm is used to perform customer segmentation, which help identify the most common/important features of existing customers
A supervised model is created to predict which individuals are most likely to convert into becoming customers in the company campaign
Project: https://github.com/typhoon1089/capstone-arvato-project
Starbucks Capstone Challenge
We process and analyse the raw data files (i.e., customer profile, offer portfolio, and customer transactions) provided by Starbucks. The main interesting points include:
Exploring data and understanding the business flow. So far, existing solutions can be found on the Internet but the authors make simple assumptions about events in customer transactions while ignoring event orders and the time instant those events happen.
Analyzing data to discover which demographic customer groups currently respond best to which offer type. At the end of data analysis, we can create new good features which are inputs of later machine learning tasks such as (i) customer segmentation and (ii) customer prediction (to determine which individual will respond to a given offer).
Project: https://github.com/Typhoon1089/capstone-starbucks-challenge
OTHER Projects
Other projects when I was a student can be found in the following pages