The Briefing


As the coaches promised before, we would be given an easier task to work on, considering we were quite stressed out in the past three days and we were going to have a trivial night in the evening, which means we would have less time to build our dashboard.

But one unique aspect about today’s challenge, though, is that we are only allowed to use Tableau version 8.3, which was released somewhere in 2017. This means no fancy features that came afterwards like viz in tooltip, set/parameter actions, LOD Expressions, etc. Also, there was no hyper file format back then, so we have to use the tde file format instead. Fortunately, the core functionalities and the interface pretty much remained the same, so it’s still very familiar to work with. I guess one real benefit of working in a much older version is to appreciate how much progress Tableau has made in the last 2-3 years.


Getting the Data


We were given a cocktail database API to start with. This allowed us to download JSON files directly from the API website. By using the JSON parse tool in Alteryx, we can easily get various kinds of information for a list of drinks. The workflow I created in Alteryx was mainly used to clean and re-structure the data, to transform it into a clean and usable format for Tableau. Here’s how the workflow looks like:



The features contained in the data include things like the drink’s name, its ingredients and respective amounts, instructions on how to make the drink, as well as links to show the drink’s thumbnail. A small number of the drinks also have a youtube url. As you can see, this data is actually very straightforward to work with.


Building the Visualization


After getting the data into tde format, I immediately started exploring it in Tableau, to get a quick idea of how the data looks like and identify any potential relationships. After this step, a really useful advice before jumping straight into dashboard is storyboarding, which basically means planning how you would lay out your dashboard, what types of charts will used, and possibly how these charts will interact each other. This process is best done using either a piece of paper or using a whiteboard. Storyboarding can save a big chunk of your time as it prevents you from just putting a bunch of random charts on your dashboard and try to make them work. This type of trial and error can turn out to be a big time waster sometimes and is best to be avoided.

The charts I decided to include in my final dashboard consist of one bar chart, one word cloud, and one web page object that has the cocktail picture embedded. This is how the dashboard looks like (which can also be accessed using this link here):



How to Use This Dashboard


This is a very straightforward dashboard to use and the main purpose is to help you make a drink based on your own selection. First choose alcoholic or non-alcoholic and you will see a list of categories under that class. Then click on a category to continue to narrow down your list of drinks. Through the word cloud you will see the most common ingredients used for this category of drinks. Last but not least, select your most desired drink from the scroll down list. And bang, you will how the drink looks like, what are the required ingredients, and instructions on how to make it. It’s time to make one of your own cocktails now!


Romy Li
Author: Romy Li

Data Consultant DSAU6