H&M - Customer Purchase Prediction
Industry
Retail
Skills
- Python
- SQL
Algorithms
- Decision Tree
- Random Forest
- Gradient Boosting
- Neural Network
- H20 AutoML
- SHAP
Introduction
H&M is a large fast-fashion retailer company with over 5,000 stores worldwide. Being a large retailer, it is important for H&M to leverage predictive analytics to improve decision-making and operational efficiencies. This project is focused towards helping the marketing team of H&M to help them understand the target customer group for marketing campaigns. The target customer group are the customers that are likely to make a purchase in the next three months. The marketing team can send a customized marketing campaign to the customers who are likely to make a purchase and bolster their revenue.
We have targeted the dataset of H&M Group (in-person store and online store) to develop a system to predict the next purchases of customers based on data from previous transactions, as well as from customer meta data. This dataset has been downloaded and imported from Kaggle using Kaggle’s json credential file and script for us to work on it.
For our goal of predicting the purchases made by customers in the next 90 days from their respective previous transactions, we set out a timeline of date ranging from 22nd March 2020 to 22nd June 2020. Historic purchasing behavior of customers were taken in account for in our data set and the dependent variable 'next_90_days_purchase' was serviced according to the models' requirements.