Using methods from Machine Learning and AI to predict income of a company

Introduction

🌟 In this analysis, we explore the intersection of machine learning and AI with business analytics. Our objective is to unveil how these technologies can accurately predict a company’s income. This journey is not just about algorithms; it’s a story of pattern recognition and strategic forecasting.

📊 Dataset Context: We use data collected by restaurant staff. It includes various attributes of the customers and their orders, offering a perfect scenario to demonstrate how machine learning can uncover valuable insights from day-to-day data.

Predicting Company Income with AI and Machine Learning

💡 The Challenge: Imagine the ability to predict a company’s income with precision. This is the core of our exploration, where we employ machine learning and AI techniques.

🎯 Our Goal: We aim to assess how accurately we can forecast actual earnings based on estimates made by company employees.

The Core Question

❓ The Main Inquiry: Can we accurately predict a customer’s spending on our products based on our employees’ estimates? This question forms the heart of our investigation and guides our analytical journey.

🔍 In the process of seeking an answer, we also unearth additional insights from the data, potentially valuable for strategic decision-making.

When the day off will save the most monay?

First, we will evaluate on which days there are the most customers so we know when taking a day off will save money. It seems that it is a good idea to take a day off during Friday. The mean income in dollars paid by individual customers is 17, while on Sunday, it is 21! So, if you have 100 customers per shift, you can earn 426 dollars more.

Click on the red dots in the figure below to see how mean income in other days are.

Who Will Consume Our Services More? Who Should We Target During Advertisement Campaigns?

We also have information about the sex of our customers. Perhaps the sex of a customer plays a role in selling our products! Let’s find out. Results show that we have a higher income from males by approximately 2 dollars. Thus, focusing advertisement campaigns on this category can lead to higher future income.

Predicting Company Income Based on Employees’ Estimates

To tackle our main question: How accurately can we predict customer spending based on employee estimates? We developed a simple statistical model to forecast actual bills from employee estimates

The model’s predictions are visually represented below, showing decent predictive ability. Specifically, it suggests that 44% of our income can be predicted by estimates from our employees! They are quite good at that. Their estimated income per customer differs from the real income by 7 dollars!