Dynamic Demand Forecast Dashboard – A Machine Learning based Quick Commerce Model
No TL;DR found
Abstract
Quick Commerce (Q-commerce) is a rapidly growing sector that delivers daily essentials within 15 – 30 minutes. Companies in this sector currently operate using a dark-store model (inventory model). However, this approach presents challenges such as high operating costs, low margins, low average orders, and inefficiencies in last-mile delivery. The aim is to propose a new operating model that reduces operating and last-mile delivery costs through machine learning. A Dynamic Demand Forecast Dashboard (DDFD) is proposed which forecasts sales for specific categories in particular locations and times. This is achieved through autoregressive integrated moving average (ARIMA) modeling and Random Forest classifier, with an accuracy of 70%. The dynamic pricing approach further allows for real-time revenue maximization, adapting to fluctuations in market conditions.