INTRODUCTION

For this project, I used a dataset created by Michael McKeown found on the CAIN website.  

The dataset includes information on deaths related to the Northern Ireland Conflict, also known as ‘The Troubles’. In June of 2009, Michael Mckeown decided to make the dataset freely available to the public for research purposes, but the database remains copyright © of Michael McKeown.

For this particular project, the focus only included those individuals who died in Northern Ireland during the conflict.

INSPIRATIONS

The first inspiration, figure 1 below, was chosen because it is a political map of Ireland. I feel as though it helps contextualize the conflict, especially to those who are unfamiliar with the history. The Republic of Ireland is represented by the national flag of Ireland and Northern Ireland by the Union Flag, as it is a part of the United Kingdom.  

 

I thought my second inspiration, figure 2 below, was interesting as a visualization.  The Cain website includes many maps made on the Northern Ireland Conflict, using the Sutton Index.  I specifically chose the visualization below because the creator decided to represent the individuals who died by name, instead of numbers, and created an abstract map of Northern Ireland from the names.    

The map creator comments,  ‘.. that the map is not meant to be viewed conventionally but it rather gives an impression of the spread and intensity of the conflict within Northern Ireland. ‘  If you click on the image it will take you to the gif of the visualization.

Lastly, I chose figure 3 below as an inspiration due to the similar subject matter, mapping the death of individual people due to the conflict. It was also created using Carto, a software providing GIS and web mapping tools, which I was initially going to use to create my visualization.

MATERIALS & METHOD

The Data

My dataset was sourced from the Cain website. The dataset is owned and created by Michael McKeown. More information on the creation of the dataset, including reasoning and methodology, can be found online from McKeown’s Post-Mortem.  

I also used a shapefile for this lab. The shapefile I used of Northern Ireland Parliamentary Constituencies was found online at data.gov.uk.

Cleaning the Data

I used OpenRefine to clean the dataset.

The first task I had to do was to remove all individuals who were killed outside of Northern Ireland from the dataset. This included the 21 individuals who died in the EU, the 115 individuals who died in the Republic of Ireland, and the 123 individuals who died in Great Britain.

The second task was matching the location column in the dataset to the shapefile. Specifically, I had to change names of the Belfast constituencies from Direction, Belfast to Belfast, Direction.  

Making the Visualization

Initially, I was working in Carto for this lab, but following a discussion about my dataset during the lab session, it became clear that Tableau would be better suited for creating the visualization I was aiming for.

VISUALIZATION & DISCUSSION

After uploading my dataset and the shapefile into Tableau I was able to create a map based demonstrating the number of individuals who died within the 18 different Parliamentary Constituencies located in Northern Ireland.

I used Tableau’s filter functionality to create 4 filters that users could interact with, creating faceted search options, including context (how the individual died), agency (what agency was responsible for the death), religion (the religion of the individual), and year (what year the individual died from 1969 to 2005).  

I then created a second map to highlight 1972, the year where the most deaths occurred during the conflict. I created two version of this. One with a screenshot within the dashboard and one with a sheet included in a dashboard figures 4 and 5 below.

Both versions have their disadvantages. Figure 4, with the screenshot, is not interactive, although the user can use the filters themselves to recreate the view.  Figure 5, the one with a sheet, is interactive, but the color key does not reflect the correct information if users start using the filters.

 

For my visualizations, I decided against the use of percentages for my filters because I felt that further removed a user from the connection of the data to the loss of a human life.  

The visualizations can be found here.

FUTURE DIRECTION

Overall, I think my final dashboards can be a bit overwhelming to use, especially for users who are not familiar with the Northern Ireland Conflict. I believe including more context in the final dashboard might help to ease this.

Moving forward I would like to spend more time with the overall design of the dashboard. I would like to spend more time with the filters and how they interact with each other and the map. I would also like to spend more time with the color key, with not only the shading scale but figuring out how to make it more obvious to users that the numbers change once filters are being used.

Looking back I am not sure if I would choose the same filters or as many filters as I did for the final map. I do not feel that the 1972 call out is working how I imagined it and I would replace the map with just text calling out the year and information.  

I need to clean up the terms used in the dataset to be more user-friendly, specifically the labeling of the agency and context filters, along with replacing the acronyms for the full name of within the agency filter.  I would also like to work with the dataset to create a column for categorizing agency affiliation (Loyalist vs Republican vs State Forces) that are available as metadata in the dataset.