It has been a few weeks since Part 1 of this series and since then I have learned lots at the Data School Down Under with still plenty to learn. One thing that I have learned was finding a story for your data. In this blog, I will go through 3 tips to help explore and tell a story with your data.

1. Prepare the Data – Make it ready for analysis

Once you have your data, it is not always going to perfect and you will find that you may need to add, remove or alter some of the data. The process of preparing the data can be iterative where you be find yourself going back to this step to make adjustments. It helps to combine this with Tip #2 below where you may want to just set aside time away to plan what you want out of this data and the story you want to tell.

I want to emphasize on the quality of the data because it will affect your end result because like when you bake a cake with expired ingredients, you can try your best to make it look nice but you wouldn’t want to serve that to anyone.

How you can improve on the quality of data – a few things to keep in mind

  • Preparing the data involves understanding what is in the data.
    Imagine column of data that is intended to indicate a numerical score but what if there is a row in the data that contains words instead of a number (eg. it says ‘Zero’ rather than 0). So it is important to understand what values are contained in each column before you proceed or else it may affect the quality of your data and analysis.
  • Case Sensitivity in fields and other slight differences
    If I had a category around colours of paint and I saw that there were values in the data with one saying ‘Purple’ and another is ‘purple”. It is vital to ensure that they match exactly so that it counts both together. If I had 50 rows with ‘Purple’, 50 rows with ‘purple’ and 75 rows with ‘Red’, if I tried to count it on Tableau, it would tell me that ‘Red’ has the most with 75 when we know that it is wrong. One way you could fix this in Tableau is to create a calculated field with something like UPPER([Colour of paint]) which will capitalise every letter so both values of purple will show as ‘PURPLE’.
  • What do the numbers mean? Check your numbers!
    It may seem simple to understand and acknowledge but still always have to stay vigilant that when you are totalling up the numbers that it is correct. For example, if I recorded the amount of coffee I drank each time I wrote a blog and I ended up drinking 2 cups of coffee for each blog and I did 4 blogs in total then it would be wrong to say I drank 16 cups in total. There could be various of reasons of why it may not tally up correctly but there are a few things you could consider such as duplicates and what is being included and excluded.

2. Ask Questions About the Data – What do I want to know?

This is about thinking what the story and purpose of your Tableau Viz is going to be. This helps determine what is going to be useful to you. You could be looking at a dataset with over 100 columns but in reality what you need actually need may just be 10. This helps with the process in Tip #1 because what is the point of cleaning something that you are going to throw out.

It helps to understand the purpose of what you are doing because when you don’t, it may lead to many rabbit holes where you want to explore everything but end up with nothing (I’ve been there plenty of times).  So it just helps to think about who would use what you are going to make and think about what is useful to them. For example, if were a coach or analyst in a professional sports team, I would want to know how well my players performed in a game and wouldn’t want to worry about how many tickets were sold at the stadium.

In essence it helps to interrogate the data and think about what is useful to know and what isn’t and what could possibly add to make it more usable.

3. Explore the Data First – Create some basic visualisations

It is easy to make assumptions and hard to not make them about data especially if you are dealing with topics that you are familiar with. For example, you would expect major cities to have a higher population than the smaller regional towns. But sometimes the data can surprise you with the results so always challenge your own assumptions and explore the data first before you start making a dashboard on Tableau.

Suggestions to explore the data on Tableau

  • A simple bar chart can go a long way. It doesn’t have to be fancy because at this stage, you are just learning about the data.
    Use only one or two fields/columns in the data and plot it to get the idea of the distribution of data
  • Time series line chart – one that could make it as part of your final product for the dashboard but also quick and simple enough get an idea of the trends
    It can be a good starting point to start your analysis due to it’s simplicity and returns a very useful amount information
  • Even just a table with numbers can help – It isn’t exactly the way some would imagine how they would use Tableau but certainly handy.
    If there many categories and sub-categories in the data that you want to explore, Tableau allows you to quickly add/remove fields you want to observe

Thanks for reading part 2!!

I hope that these 3 tips will help you understand more about the process before creating a dashboard to show off. Please watch out for my next part for my Project Starter series and feel free to check out my other series, Don’t Be Me, where I will post the second entry for it soon. Thanks again for reading!

Kier Bituin
Author: Kier Bituin