Food Hire Forecast

📋 Problem Statement:

The pandemic has forced all public places to shut down. As a result, many businesses, such as eateries, struggle to cover financial costs to continue with their services. To make up for severe financial debt, restaurants resort to inflating their menu prices. The problem with this action is that restaurants are attracting fewer customers. Customers are their main source of income, and part of it is used to cover financial costs. Human resources are also a big factor for costs that the company faces, and after the multiple layoffs during the pandemic, companies are starting to recover.

Businesses often do not efficiently allocate their resources, such as staff. They might overhire for slower business periods, or not hire enough staff during peak times.

❣️ Our Mission:

Our mission is to create a model that allocates an appropriate number of staff on a given week/day/holiday/hour based on the data sets. That is, if more customers show up in a given time frame, the restaurant can use the data from the model to decide to hire more or less personnel during those peak times.

⚙️How We Built It:

We found the frequency of the bills using the Python Pandas + NumPy libraries to gain insights on trends. Then, we trained a model with scikit-learn’s Random Forest Regressor using those columns to predict an average number of customers that the restaurant could expect on any day when given a user inputs a concept, a date, and a time.

🤔 Challenges We Ran Into:

Formatting the data was a big challenge as we had to work with a lot of categorical data such as dates, cities, and concepts. Training the model was also a challenge because of the large

✈️ What's next:

To improve the accessibility for viewers, we would use UI interfaces. One of the software we could consider is using Streamlit. We would like to make the visualization of data more ready for users when they want to investigate more into a particular city.

With more data we can provide more specific insights for each specific restaurant instead of a city/concept in general and create a dashboard to display those results for businesses.

Built With

• matplotlib
• numpy
• pandas
• python
• scikit-learn