British Airways (BA) is the flag carrier airline of the United Kingdom (UK). Every day, thousands of BA flights arrive to and depart from the UK, carrying customers across the world. Whether it’s for holidays, work or any other reason, the end-to-end process of scheduling, planning, boarding, fuelling, transporting, landing, and continuously running flights on time, efficiently and with top-class customer service is a huge task with many highly important responsibilities.
As a data scientist at BA, it will be your job to apply your analytical skills to influence real life multi-million-pound decisions from day one, making a tangible impact on the business as your recommendations, tools and models drive key business decisions, reduce costs and increase revenue.
Customers who book a flight with BA will experience many interaction points with the BA brand. Understanding a customer's feelings, needs, and feedback is crucial for any business, including BA.
Once you have your dataset, you need to prepare it. The data will be very messy and contain purely text. You will need to perform data cleaning in order to prepare the data for analysis. When the data is clean, you should perform your own analysis to uncover some insights. As a starting point, you could look at topic modelling, sentiment analysis or wordclouds to provide some insight into the content of the reviews. It is recommended to complete this task using Python, however, you can use any tool that you wish. You can use some of the documentation websites provided in the Resources section below to analyse the data.
Your manager would like you to summarise your findings within a single PowerPoint slide, so that they can present the results at the next board meeting. You should create visualisations and metrics to include within this slide, as well as clear and concise explanations in order to quickly provide the key points from your analysis.
First, spend some time exploring the dataset in the “Getting Started” Jupyter Notebook provided in the Resources section below to understand the different columns and some basic statistics of the dataset. Then, you should consider how to prepare the dataset for a predictive model.
When your data is ready for modelling, you should train a machine learning model to be able to predict the target outcome, which is a customer making a booking. For this task, you should use an algorithm that easily allows you to output information about how each variable within the model contributes to its predictive power. For example, a RandomForest is very good for this purpose.
After training your model, you should evaluate how well it performed by conducting cross-validation and outputting appropriate evaluation metrics. Furthermore, you should create a visualisation to interpret how each variable contributed to the model. Finally, you should summarise your findings in a single slide to be sent to your manager.
- Python - Programming Language
- Jupyter Notebooks - Open-source web application that allows data scientists to create and share documents that integrate live code, equations, computational output, visualizations, and other multimedia resources.
- Anaconda - Local environment for practice
- GitHub - Repository for storing all files