DSAU25’s first client project was with a company who gathers flight and tourism data, and on a rain filled Friday, we presented what we had found in the dataset over the course of a week of exploration. We first got a glimpse at a sample of the dataset the week prior and were given tips on how to start brainstorming through such meetings and collaboration efforts like ‘Scrums’ and ‘Sprint Plannings’.

With Jay at the helm of our first project as the ‘Scrum Master’, we divided up the tasks in terms of data cleaning and preparation, and the interpretation and visualization of the data. Our first step, which was performed by Mika and Daniel, involved using Alteryx to clean the provided data and join it into a dataset that included the full airport and country names, as the original dataset only contained their abbreviated codes. A Union tool was then used to bring the datasets together.

What followed was creating the exact points as to where destination and origin countries/airports were using the provided latitude and longitude points in the dataset. The exact locations were initially in columns with the word ‘POINT’ at the start, followed by the latitude and longitude in brackets. These were cleaned and moved into separate columns for origin, destination, and the two stopover locations ‘Via 1’ and ‘Via 2’.

The data was then split into two different variations – one was the ‘Passenger Data’, which held information on the route, when it departed, and the passenger traffic numbers (PAX) that were estimated to be on the flight. The other dataset created was the ‘Route Data’, which contained information on the origin and destination locations, the stopover locations, what the operating airline was, and the direct and total distances required for the given flight. A couple things in the data had to be fixed as we worked on the project, such as the transposing of the PAX numbers for each class into rows from columns, and the fixing of geographic locations, be it the inconsistent naming or having the wrong coordinates provided.

The next step was to create the data visualization dashboards, and this task fell primarily on myself and Prerana, though the team collaborated together in an efficient manner, and everyone chipped in with creating different visualizations. We had an overview page which contained our overall insights on air travel based on the dataset, a dashboard on the high-value customers, and two pages on what we believed to be the underserviced markets (the data has been edited out due to privacy reasons).

The presentation went as well as it could for a first client project, especially since we only have a week to understand the data and create something from it each time we do these projects. Being a team member was a great experience for our first project, and it gave me a chance to put to use everything we’ve been learning in Tableau. I made use of a number of features we have learnt over the past 6 weeks, including but not limited to table calculations, sets, parameters, and spatial mapping.

The biggest thing I learnt was how exactly a project works. In a team, you don’t have to do everything, but everything that needs to be done does need to be accounted for. It was interesting to work on my own part while still making sure I understood the data and the requirements of the project from a wholistic view. I really enjoyed the experience of brainstorming and collaborating with others to find the best way possible to analyze and create insights from the data. There were definitely a lot of aspects about project work that I learnt throughout this project week that I will be able to use in future projects.


Nicholas Seah
Author: Nicholas Seah