Data Visualization

I Guess I’ll Have to Change My Plan: Data Visualization

The Final Product!

At this moment, I would not call my experience with Tableau a rousing success.  I decided that I would use data that will be useful to my thesis – a representation of where the US government recorded US ships being taken when they were captured by the French.  More specifically, I want to show the reality that US government charts showed at the time – that the vast majority of seized ships were being taken to St. Domingue – present day Haiti.  Even some of the ships that the government designated in a more general list of captures were also headed to St. Domingue, they just had specific ports within the colony listed instead of the name of the colony itself, and so were not counted as being taken there.

This map does roughly what I wanted it to do.  Using the data in a map form shows more than the tables that the US government used for this type of data.  This is partially because the data tables could be published en masse in a book that was just tables: a needle in a needle stack.  While the map does not give as much data as these tables (such as telling what kind of cargo the ship carried or what happened to it after it was stopped), it does focus the attention on the geographic concentration of the French actions, which is not as easy to see in the tables.

The process of creating the map was not easy – once I got the data inputted, all of the dots on the map were different sizes, but none of them proportional.  I changed types of display – I tried a table, and a bar graph, and two types of map, and it did not accurately display the scale of what I was trying to describe. Throughout this process, I followed three different help desk articles, reworked the data upwards of 10 times, but it still showed the dot in Cuba, where three ships were taken, as the same size as the St. Domingue dot, where 45 ships were taken.  I finally got it to work by redoing all of my data into a frequency table and removing all of my identifiable data about the ships – which is not ideal, since it will make the chart only useful as a visual aid, and not useful to me in terms of discerning any patterns in what types of ships were targeted.

Once I had the map that I was satisfied with, I had one final hurdle – posting it to this blog.  This task was made more difficult by the fact that as soon as I hit save on Tableau, every other program on my computer froze and shut down.  In a particularly weird turn of events, the internet browser that I use stopped functioning, then I forced it to quit, but all of the windows stayed open for another 15 minutes.  They were not clickable, they just sat there, haunting the computer.  Once I restarted the browser, instead of just reopening normally, it reopened more un-clickable windows.  I don’t know that it was caused by Tableau, but it has never happened before.

Ultimately, I think that Tableau could be a very good program to have on an office computer – it is useful to be able to create attractive graphics to make your point.  However, my main conclusion about data visualization in this project is not necessarily about the technology.  I think that the type of data I used does not necessarily lend itself to the types of maps that I saw as available on Tableau, and I think that the interpretation I want to use is better accomplished by a written narrative.  In other words, it is as important to know when not to use data visualization is it is to know when it can be helpful.