Every year, hundreds of journalists around the world are physically and verbally attacked, jailed, and killed. While in the United States we enjoy decent press freedom (45th out of 180 on the 2018 Reporters without Borders World Press Freedom Index), this is historically not the case in many other countries. Therefore, for my second visualization lab, I chose to work with data on the number of journalists killed worldwide from 1992 to present. This data was compiled by the Committee to Protect Journalists.
Critique of Previous Visualizations
The visualizations that I found are not about the same topic, but a related topic: the jailing of journalists worldwide. Unfortunately I did not find a visualization that I believed to be particularly successful, but feel I can learn from the mistakes of two of them.
While this visualization (Fig. 1) makes sense within the context of the post where it was found, if it is removed from the post, it does not contain enough context to be understood. Additionally, not all readers are familiar with the abbreviations of countries, and their placement on the visualization only vaguely correlated with their place in physical reality. Unless you are dealing with a small geographic area that most of your primary audience is acutely familiar with, the full names of places should be spelled out.
This CARTO visualization of the Islamic States’ effect on the press is a visually striking one, but I don’t find it to be particularly useful for finding more specific information. Each dot on the map contains the name and information of a journalist who was killed by the IS. While it does convey the idea that the IS has killed many journalists in the last six years, the overlapping dots and fast-moving time lapse makes it difficult to see information about the dots that appear early on in the visualization but get covered up by later dots. It is also hard to tell the dots apart and a group of dots turns into one large mass. While I do think maps are generally most effective for this topic, this visualization may need some tweaks to be as useful as possible to researchers and everyday readers alike.
For this visualization, I used Tableau Public. I downloaded and ran the program on my desktop, and later uploaded the visualization to the Tableau Public site to save it. After finding a dataset in the .csv format from the Committee to Protect Journalists and ensuring that it was clean enough to use, I uploaded it onto my desktop application. Tableau generally recognizes the kind of data that you have uploaded (years, countries, etc), which makes using the data relatively easy.
I chose to use all of the data available on journalists killed from the Committee to Protect Journalists. I could have perhaps limited it to the last 10 years (or since the beginning of a specific conflict), but because this was exploratory, I used the entire dataset. I am not an expert on international journalist killings, so most of my visualizations were experiments, to see if interesting, striking relationships would emerge out of different combinations of categories.
Because I did not have an explicitly quantitative row/column, I worked a lot with Tableau’s “Number of Records” function, which counts the number of times a record of a certain category appears under another category (See Fig. 3 for an example).
I created eight visualizations and one dashboard (including four of the visualizations) in Tableau.
For the dashboard, I chose the four visualizations that were most compact and easy to read. The visualization that I found to be the most successful was the map. The map provides approximate numbers for every country and can also give you information about regional trends. I don’t think it’s a perfect visualization, as it only provides approximates, but it is easier to interpret than some of the others.
Visualizations that listed all countries individually were often too large and required scrolling to read in their entirety (See Fig. 5-8). While I do think that these visualizations would be useful for researchers and could be useful for lay readers if they were able to be displayed on one page (perhaps in a newspaper), generally the more compact visualizations can give you more of an impression of the situation overall.
To view all of the visualizations, see the following links:
- Total Journalists Killed, 1992-2019
- Total Journalists Killed: Local & Foreign, 1992-2019
- Total Journalists Killed By Country, 1992-2019
- Total Journalists Tortured By Country, 1992-2019
- Total Journalists Killed by Country, 1992-2019
- Type of Death by Country (Totals), 1992-2019
- Total Journalists Killed By Country, 1992-2019
- Total Journalists Killed by Gender, 1992-2019
A challenge that I encountered while working with the dataset was my limited knowledge about the subject. Perhaps there is a benefit to working with a dataset about a topic that you are not familiar with, as you are less likely to find specific patterns that you are already searching for, but if there is previous knowledge about the topic, you are also better able to know what factors to look for or what parts of the data may yield a meaningful visualization. I would have preferred to work with a dataset about the jailing of journalists worldwide as it is something that I have been following in Turkey since 2012. Unfortunately I was only able to find datasets on this topic from the Committee to Protect Journalists for 2017 and 2018.
I also think this visualization project could benefit from normalization of the data by including either population of each country listed, or, more ideally, including the number of journalists in each country. Giving a percentage of those killed or a number compared to the total number may be more impactful when telling the story of the dangers the journalists face to do their work.
I do think Tableau Public is a perfectly suitable tool to investigate and visualize this topic, but I might prefer a tool with a greater variety of mapping capabilities. Because there are regional variations in journalist killings (likely due to violent conflicts), maps would have provided a more clear representation of the situation and would have been able to accommodate the other factors that I incorporated into my visualizations (total over time per country, gender, number tortured, etc).