FOODHUB ORDER ANALYSIS

Data OverviewUnivariate AnalysisMultivariate AnalysisData Visualisation

CONTEXT AND OBJECTIVE

In the framework of the MIT Applied Data Science Program, an exploratory data analysis (EDA) case study had to be realized for the food aggregator company FoodHub, based in New-York. The company offers a convenient solution through its app, which handles the entire process from order placement to delivery, therefore addressing the growing demand from busy students and professionals. FoodHub earns revenue by taking a fixed margin from each delivery order.

The company has collected data on customer orders through its online portal. The objective was to provide insights to improve the business and enhance customer experience.

WHAT WAS DONE

An EDA was performed on the data provided by FoodHub, to extract actionable insights and recommendations. More specifically, a data overview was realized, followed by univariate and multvariate analysis.

The seaborn Python package was used for data visualisation.

From this analysis, 9 key observations could be drawn, and 8 business recommendations were formulated.

Finding the popular types of cuisine

A count plot was used to visualise the number of orders for each type of cuisine. It turned out that, out of the 14 types offered, 4 were by far the most popular. Therefore, one recommendation was to expand the network by focusing on restaurants serving these popular cuisines.

Order costs distribution

Using a combination of histogram and boxplot, it could be noticed that the distribution was right-skewed, with quite a few orders above $20. Further investigations showed that these orders represented the main source of net revenue, and one recommendation to business was to encourage restaurants to offer bundled deals with extra items (e.g., desserts) at a reduced price, to increase the average order cost.