giftcard

Problem Statement

Given the giftcard purchase data from 01-01-2021 to 01-14-2022, predict the value in monthly total purchases for the rest of 2022.

What are the highlights from this project?

This data analysis project showcases a time series analysis and forecast modeling. By processing the data and using the Prophet model to build a time series forecast model, I supported my stakeholders in anticipating future performance, which is key to optimize resource allocation and support strategic planning. Despite initial data challenges in the model selection, my efforts resulted in accurate predictions of total purchases (in denomination) for up to 4 currencies, providing a remarkable lift to the Finance team.

What to expect in this repository?

1. Order_items.csv, a dataset which contains information on individual gift card, includes variables such as unique id, created date, product currency, sender name, delivery type, recipient name, recipient email, recipient address, and more. The data ranges from 01-01-2021 to 01-14-2022.

2. Digital_card_orders.csv, the parent record of Order_items.csv, includes data such as unique id, created date, order_id, status, country and language. All data ranges from 01-01-2021 to 01-14-2022.

3. A jupyter notebook in storing all the scripts and relevant charts for data manipulation, data pre-processing, time series decomposition and analysis, forecast modeling, model fine-tuning, as well as the end result.

What is the source of the data?

The data files are provided by a company who specializes in providing technology solutions in revolutionalizing the digital gift card industry.