Analysis of Consumer Behaviour in Food Delivery Platforms

Abstract

During recent years online food delivery applications like Zomato, Swiggy and others became a routine for many people. Nowadays, it is even easier and more comfortable for the young generation like students ordering food online instead of going out and dining somewhere. Therefore, there has been a generation of a lot of information about people and their behaviour that can help to find patterns and understand the target audience better.

To avoid using someone else's dataset, we developed and distributed our own Google form and obtained 462 responses from actual customers. Thus, we had a chance to get behavioural data and see actual trends in our field. We started working on the task by performing preprocessing and analysing of the received data using Python. We managed to visualize various trends related to expenditure per order, ordering frequency, time, platform preference, etc. Moreover, we developed a machine learning algorithm using Decision Tree classification.

As a result, we found that people preferred moderate expenditures, ordered mostly at night time, and preferred to use Zomato service. Our predictive model reached only approximately 21.7% accuracy and proved the complex nature of humans' decision-making process.