Hello and welcome to day 3 of Dashboard Week!

 

After we got our presentations for yesterday’s dashboards out of the way in the morning, we got cracking on our next project. The data for today is related to Star Trek.

 

Rather than being given a flat file (a.k.a. a single sheet or table in a file or database), we are given a link to an air-table.

 

Something Doesn’t Add Up

Curiously, the suggested aim for today is to provide a way to “keep a track of all things Star Trek”. It seems an odd request given that the link provided is the answer!

 

I assumed that I should not simply embed the webpage in a Tableau dashboard and call it a day ha-ha. I do feel that although that would complete the brief perfectly, it would defeat the purpose of Dashboard Week. (Although it certainly would have sped things up for me today).

 

Searching for a Topic

I therefore decide to see if any insights could be gleaned from the data provided. Unfortunately, given the nature of the data, it is extremely difficult to analyse and provide any kind of interesting insights within such a short timeframe.

 

The air-table had 10 sheets of data on Star Trek topics including things such as series, fan items, games, podcasts etc. I decided that I was most interested in series, movies, games, official books and comics and ignored the other tables.

 

Hurdles

Initially I wanted to try do some kind of machine learning to analyse the titles and see if I could group them into any kind of logical themes. I also wasted the rest of my morning looking for a data set of names and genders in a hope to be able to feed the writer names into an algorithm and find out their gender.

 

After struggling with Python, JSON and an API, it was fast approaching lunch time. I realised that I needed to abandon my current course of action and start playing with charts in Tableau to see what I could visualise.

 

Moving Forward

Using the formula tool in Alteryx, I added a “Media Type” to each sheet as well as a column for counting (because there is no column 100% full of data and consistent between all sheets). Next, I used the selection tool to change some column titles to be consistent across tables. I then used the union tool to join all the data together so that I could make comparisons in Tableau. Finally, I exported the data, ready for Tableau.

 

Deciding on a direction

The hardest part of today was finding anything of interest within the data. I eventually decided to go with looking at gender (for authors as well as if gender is referenced in the title of a work). In a compromise on my want to analyse topic trends within the data using machine learning, I settle on a word cloud. Back in Alteryx, I split the titles into separate words, one per row, and then transpose them into a column. I then use this column to create the visual in Tableau.

 

The final result

For my visualisation, I decided to split this into two dashboards. One showing more general data as an introduction. I then decided to have a separate dashboard for the gender analysis.

As usual, I would have liked to have done more. But I am happy that I found a different angle to explore – one that no one else considered. The design could be more polished, but the dark colour scheme and formatting I have applied fit the theme of the topic, which is important.


Dashboard week - day 3

Reflections for tomorrow…

Today was more painful than it should have been. I needed to stop earlier when I was stuck on my machine learning trail. I am looking forward to tomorrow to see what others have created.

 

Emma Wishart
Author: Emma Wishart