Tableau and Battle Deaths Dataset
For the data exploration assignment, I used Tableau, data visualization software and the Battle-related deaths dataset v5 2002-2007 the Uppsala Conflict Data Program (UCDP). This and other UCDP datasets are available online at: http://www.pcr.uu.se/research/UCDP/data_and_publications/datasets.htm.
This is one of my visualizations charting battle deaths (y axis) over time from 2002 through 2007. Each different bar shows the sum of estimated battle deaths in that location for that year. I created a filter to change the color of any battle where the U.S. was involved with a pretty striking result — the is reported as having been involved in the battle locations with the highest battle deaths in 2003-2007.
Battle-related Deaths per Battle location (2002-2007), US Involvement highlighted in orange
Tableau
Our class is licensed to use Tableau at no cost during the term. Usually, the standard software is $900 and the deluxe is $1800 — pretty steep. I was happy with it for this assignment. It automatically parses data from many formats. I loaded mine as an excel file. The parsing was done well — Tableau assigned fields as nominal and ordinal correctly. I had to do a fair amount of reformatting of the geographic fields to make them recognizable to the mapping tool. Tableau has a mapping tool! If you have geographic data, you can plot associated ordinal data on a map using a variety of tools.
There’s a new free/public version of Tableau available for download and also a public competition for visualizations made using this tool.
Tool download: http://www.tableausoftware.com/public/download
ReadWriteWeb and Tableau Public Challenge info: http://www.tableausoftware.com/public/viz-contest
Wow, really interesting data sets. I threw some of this stuff into ArcMap (also a pricey tool) and got this: http://www.flickr.com/photos/benchun/4433039016/
Great viz, Becky. It’s clear that US involvement changes the severity of a conflict greatly. I wonder if you could also look at this by number of combatants involved and type of weaponry used– is the difference due to more soldiers being deployed, greater battlefield technology or something else?
You could also download the free Tableau Public tool and publish this interactively, or publish directly to the web from Tableau Desktop 5.1 or later.
Nice, Ben. Do you like ArcMap?
Thanks for your comment Ellie — this data might be available in another data set and would surely be interesting to visualize. Type of weaponry and the participant’s national military spending could be interesting to visualize also. I’ll check out the free tools when my license expires in a few months!
I don’t really have anything to compare ArcMap to — it’s my first experience with GIS software. I have to say, hooking up geographic concepts to a relational database in a deep way plus putting some cartographic / layout tools on top is pretty compelling. And I haven’t really dipped into building models or anything advanced yet.
In learning ArcMap, it really helps to be familiar with database concepts (writing a query, understanding joins, etc) but it also just takes some time to learn the quirks of what everything is called and where the different abstractions are (and aren’t) made. Someday we should have a viz nerd fest where everyone gives a 5 minute presentation about some crazy tool and what it does.
One thing that I was wondering, after seeing your graph, is what are the average number of deaths in these conflicts (US-involved, US-not-involved, and total)? It would be cool to see horizontal lines showing those averages — or maybe just a summary table with these stats. It’s always nice to quantify the things we get from a visual first impression.