This repository contains coursework for the Market and Economic Research and Analysis course in the MS Applied Business Analytics program at Boston University. Team Members: Shimony Agrawal, Ziwei Cui, Shreya Jain, Wuming Zhang
The dataset “Instacart Market Basket Analysis” has been obtained from Kaggle. The dataset published in 2017 contains 3 million online orders from more than 200,000 Instacart users. There are 49,688 products arranged in 134 aisles, under 21 different departments. The report deals with the analysis of the Instacart market, conducted by one of its partnering grocery chains – Star market. Exploratory data analysis and Market Basket Analysis was conducted on 3 million grocery store orders on Instacart to make predictions of basket sizes.
- How can Star Market improve its performance on Instacart using better strategies related to inventory management, promotional messages, online content placement and recommendations on Instacart? • What are Instacart’s most ordered products? • What are the priority products and most reordered products on Instacart? • When do consumers place the maximum orders? • What itemsets do consumers order the most frequently and the top association rules between the ordered products?
- We conducted Exploratory Data Visualization to identify the top 15 products ordered on instagram, distinction between organic and non-organic products, most reordered products, high priority products as well as the maximum products ordered by day of week and hour of the day.
- Using Association Rule Mining / Market Basket Analysis, we found the top 20 frequent itemsets with a support of 0.03 and confidence of 0.06. This gave us the most frequently ordered items and helped in assessing in the buying behaviour.
Based on this analysis, Instacart can improve its inventory management, promotional messages, content placed as well as maintain an effective recommendation system.