Data Storytelling Studio 2020

Maria Popova has always been one of my favorite writers and I found this graphic on her blog brainpickings.com. The graphic shows sleep habits and literary output for various famous writers, ordered starting with Honore de Balzac who woke up at 1am and ending with Charles Bukowski who would get up at 12pm. It packs a good amount of information into a small icon for each writer – name, lifespan, wake-up time, number of books or collections published separate by category, and awards won. A lot of the information is encoded by colors, mini bar charts and symbols that are unlabeled, which saves space but does require the reader to look at and understand the key provided. All the information included in the chart seems to have been deliberate and relevant to the question for example, writer lifespan was included for context since “length of a writing career influences the volume of literary output” as explained at the top of the chart. The chart leads the eye very clearly, with a line going through the authors presented and a meaningful ordering. The illustrations of the authors make the chart more engaging and remind the reader that there are individual humans behind the data. There’s also visual reinforcement of one of the major datasets – wake-up time – since different times appear at different points around the circle enclosing each writer’s portrait. This does mean there is a good amount of blank space in the figure since each writer’s information is centered for consistency but there needs to be a wide spacing to handle the bar charts of productivity pointing in different directions. I do think this chart is not as effective seen on a large scale – it is meant to be viewed as a poster, but it’s hard to see anything viewing the entire chart in a browser window for example. It’s harder to see overall trends since the individual information is too small when the chart is viewed as a whole. It’s only when you zoom in on a specific writer that the data starts to come into focus. This chart seems to be aimed at people who are interested in literature and potentially those who are interested in optimizing their daily routines for productivity. For people who are willing to spend some time and zoom in on the writers – or maybe just look at their favorite writers – this chart is a fun look into how famous writers lived their lives and a chance to consider how much wake-up time could affect literary productivity.

The graphic shows the Iowa Caucus results that were last updated Feb. 6, 2020, 12:11 a.m. ET.

The aim of the graphic is to give details about how the candidates are doing in different parts of the state at the most recent count.  I believe it is effectively showing the key areas are for specific candidates. It is slightly unclear who has an overall majority however the graphics are ordered with the candidate with the most votes showing first, but this is secondary to the bright pictures that show the spread over the state.

The audience is people who pay attention to politics, as people who don’t are likely to only read headlines and leaderboards, especially before the votes have been completely counted.

The interesting part about live-updating graphics is that they need to tell a story even as the data used is continuously changing. When choosing scales on static data, it is easier to select those which highlight specific points or accurately give the details required.

In this graphic, I presume that there is a ratio to translate the number of votes to circle size. It is possible that they used estimates of voter turnout to determine the ratio before results came in, in which case there is a secondary source of data that went into creating this graphic. If this is the case, choosing ratios too high would create circles that are too small and similar in size to be meaningful. Choosing a ratio too small would create circles that are too large and overlapping. On the other hand, if the ratio was dynamic, a person continuously checking the graphic could see changes that imply that candidates have lost/gained votes as the ratio adapts to the data.

http://maps.sasaki.com/visualizations/homelessness/

This is an online interactive tool developed by Sasaki for understanding the homeless crisis in the US, which I recently saw presented by its creator, Gretchen Keillor . Relying on data from January 2015 where volunteers manually counted over 500,000 homeless, the project begins by representing every 5 homeless people as one dot and showing where they reside geographically.

The website the quickly turns control over to the user through a simple options panel of drop down menus at the left side of the page, which allow other information besides location to be represented on the map.

There are over 30 parameters associated with homelessness that can be overlaid onto the map, ranging from max and min temperatures, to per capital educational spending, to the margin of Barack Obama’s 2012 presidential win, to the per capita number of homeless bed available. As interesting as seeing how these different data points overlap is, the interface then goes a step further. Although it’s helpful to see how these things map geographically, it’s difficult to compare amounts when they’re represented as dots strewn across a map. For this reason, the menu to the left allows the user to then rearrange the dots into bar charts, scatter plots, grids, etc, while adding organizational layers like region, children v. adults, or bed types.

The graphs generated by the user’s drop-down selections then allow the user to scroll over them in order to get labels about the groups represented, like region if the graph is organized that way, as is shown in the circle pack diagram above.

Finally, the interface includes a “tell me a story” option, where the data drop downs are turned off, and the user is led through one specific narrative, including small blurbs explaining the take-aways from each digital representation created.

I think the major success of this interface is that it allows interaction and exploration from users without showing any large statistics or numbers, relying entirely on the visuals to give impressions of relationships between parameters. In a way, however, this is also it’s greatest problem, as some visualizations generated are extremely difficult to gain intelligible findings from, and the user can easily become lost changing parameters and representation style without any meaningful understanding of what the dots zooming around and rearranging actually implies. Even the “Tell me a story” option fails to provide an actual narrative, instead simply making a few choices about which unrelated homeless stats to show. Additionally, by representing 5 homeless as 1 dot, and then allowing the dots to be moved off the map and into graphs, it’s relatively dehumanizing, as I personally quickly lose sight of the fact that each dot is 5 people.

Overall, I like this project, and think it serves as a good model for how to engage an audience in a particular problem, even if I wonder if the final result could have been more successful if the users had been given slightly less control.

https://bostonopioid.github.io/discarded-needle/index.html

I first explored Discarded Needles last year, and since then its impact has stayed with me and to an extent changed how I view the city of Boston. It combines its primary dataset with a series of gripping interviews and images that really enhance the entire narrative. All in all, I believe this is an excellent example of data storytelling.

The story aims to examine the rise of the opiate crisis in Boston using the dataset of citizen reports of needles on the ground. I imagine this data was taken from the local 311 office, and perhaps it could be taken through the Freedom of Information Act. The story opens with the total amount of needles reported in large bold letters, a staggering amount that draws in the reader.

From here, the writers weave together several interviews alongside heavier data analysis to weave their story.

This bar graph does alright at showing the magnitude of the problem over time, but I think even more revealing is when each data point is laid over the city of Boston based on time and geographic location.

Apart from the data, the background of the story is filled with pictures of needles found around Boston, decrepit areas, and other elements of the issue. The most pivotal piece of media in this story is the final picture, drawn by a first grader talking about the needles they find around their school. 

I believe the goal of this story is to highlight the opioid epidemic in and around Boston, an issue that is obviously heavily affecting the whole nation but can be completely overlooked locally by Boston residents (as in my case). The writers do not assert a path forward or lay out clear next steps, instead they include many instances where the volunteers and people currently involved feel futile and ignored by the local government. In this way, I believe this piece is also designed to raise awareness of the government’s apathy toward this subject. I believe the audience for this piece is local Bostonians. Many of the locations, including “Methadone Mile,” “the Long Island Recovery Center,” and “the Orchard Garden School” are not explained in a way that would give context to those outside Boston. On top of this, the lack of a real call to action at the end of the story reveals that the authors intended for this story to make an impact on the local level rather than a national one.

Source: The Economist

I saw this data representation called “Map of Misery” on the Economist recently, which shows the change in suicide rates in the United States.

Using county-level CDC (Centers for Disease Control and Prevention) data on the nearly half a million 25- to 64-year-old Americans who committed suicide between 1999 and 2016, the scientists calculated the expected number of suicides for each city. Then compared these expected values calculated from the past with the observed suicide rates to show whether the suicide rates increased or decreased.

The goal of this data presentation is to alarm people about the rising suicide rates in the U.S. The Economist article I read mentions that more than 48,000 Americans had taken their own lives in 2018, equivalent to 14.2 deaths per 100,000 population. This makes suicide the tenth-biggest cause of death in the United States—deadlier than traffic accidents and homicide. With these facts and the data presentation, the author of the article wants to raise concern about the climbing suicide rates and show the geographical distribution.

Another goal of the data presentation is to draw attention to correlations between suicide rates and geographical factors, such as ease of access to guns, deprivation, opportunities for social interactions, loneliness.

I think the “Map of Misery” does create a depressive mood for the audience, yet fails to have an alarming effect. If the use of colors were reversed and the higher suicide rates were depicted in red, the map could look more alarming and be more effective in raising an immediate concern.