The language conservancy reported that, “For the last 400 years, Native American languages across the United States have been dying out. Over 200 have become extinct.” With the rising and long overdue normalization of “land acknowledgement,” I thought an interesting way to build upon my previous project (where I focused on mapping critically endangered languages), would be to represent Indigenous populations across the United States and explore the correlation between population and language vulnerability.
How is this different than my previous project?
I feel like my previous lab was visually cluttered and I wanted to use this opportunity to create visuals that take the user on an exploration through the data, as opposed to just confronting them with it as I did in my previous lab.
Dataset & Tools
I used https://data.census.gov/table?q=American+Indian+and+Alaska+Native&tid=DECENNIALPL2020.P1 and https://www.theguardian.com/news/datablog/2011/apr/15/language-extinct-endangered#data to create my visuals. I used OpenRefine and Tableau software to clean, trim, and work with my data. I intended to use QGIS however I found Tableau allowed for more flexibility and felt more familiar than QGIS. I used figma to create the final visual.
Methods and Process
I enjoyed using the same dataset as in my previous lab but this time experimenting in Tableau, rather than QGIS. Open refine helped me to remove columns and comb through my data more thoroughly than I had in the last lab.
Visualizations and Interpretation
Link to Tableau Interactive Visualizations
In the first visualization, I chose to create a tree map of endangered languages across the globe. My choice reflected my overall aim to start with the larger picture and narrow down the data exploration as the user clicks through. I couldn’t get tableau to display more text than it shows in the image however when one hovers over the boxes, the tooltip displays the information necessary to interpret the data.
The packed bubbles visuals represent information that is able to be located in the previous visual, however, I found it helpful to create two representations of both ‘severely’ (representing languages with at least 700,000 speakers or less) and ‘critically’ (representing languages with at least 51,000 speakers or less) endangered languages to allow users to more easily explore the more at-risk languages.
As I narrowed the scope to the United States, I thought it’d be helpful to see the languages displayed in alphabetical order. I wanted to make this searchable for a few different ideas (of the intentions of the person using this); for this visual (Figure 4) I had the idea that someone could potentially be looking for a specific language and perhaps knew what letter it began with or how it sounds but not how it’s spelled. Figure 5 helps comprehend the more less endangered languages, highlighting some familiar names such as Navajo and Sioux. Additionally, I changed the colors to brighter greens (& blues and pinks) to indicate the shift and focus on the U.S.
These maps allow those viewing them to picture where endangered (6) and extinct (7, below) languages in the United States are or were spoken. I chose a darker map to be able to keep the colors used in the previous visuals and have them look more attractive than I figured they would on a lighter map. When I mapped extinct languages in the U.S. I opted for a lighter map.
Figures 8 and 9 reflect the indigenous population by state; figure 8 represents those self-identifying as ‘American Indian’ or ‘Alaskan Native’ and figure 9 represents those who self-identify as ‘Native Hawaiian’ or ‘Pacific Islander.’ I chose to use an area graph rather than a bar graph because I thought the size comparison between peaks was more striking than the bar graph visual.
Finally, I created this visual in Figma for reasons I will address in my reflections portion:
User Testing
I asked two of my peers, one who is a data analyst and one who has no relation to this field, to interact with and report conclusions drawn based on the visuals, absent of any additional guidance or context. I wanted to see if the labels, headings, legends, and visuals provided enough direction to the user. I asked them afterwards if they had any questions or if they were unsure about any visuals. I did change and eliminate a few visuals based on their feedback; initially I had the first visual represented as ‘pack bubbles’ and the users supplied the critique that the lead visual should look a bit more ‘organized’ rather than scattered.
Reflections
I compared the visuals in figures 8 & 9 with the visuals in figure 6 & 7 and as shown in both visuals, California is the state with the highest number of self-identifying indigenous people (fig 8&9) and has the highest concentration of critically endangered and extinct languages. This is similar to Arizona and New Mexico who also boast larger populations of Indigenous people.
My goal for the final selection of visuals was to have a map that showed both population by state and languages’ degree of endangerment. I couldn’t, however, figure out how to do this.
Here are some things I tried:
Finding different, but similar, data. I couldn’t. I thought this https://www.census.gov/tribal/ tool might allow me to map something similar to my original idea. The links on the site did not lead me to a clear explanation or dataset which only furthered my confusion.
I tried manually re-writing the data as a geojson file, one of the users I asked to participate in my ‘testing’ phase suggested I do this. I used the resource geojson.io and this website: https://observablehq.com/@rdmurphy/u-s-state-bounding-boxes to write this code, however I did something incorrect when it came to nesting the data and despite trying a json reformatting tool, I couldn’t seem to get it to work in tableau.
I considered pivoting the data, however I already had the visuals labeled fig 8 & 9, and I was worried that if pivoting the data didn’t work that I’d lose the figures above.
Despite this short coming, I liked the visuals and the story they allow the user to piece together.
I think compared to my previous work this is more visually appealing and I think I organized the data a bit better in Open Refine.
Eliminating the world languages bit was useful to furthering my project goals and allowing the data to be more clear.
Another challenge was attempting to ensure accuracy as far as location went because in my previous project in QGIS I found it difficult to tell and I think I misplaced some dots while trying to relabel them. An example of this is Lower Chehalis:
Citations
(2023, April 25). The Language Conservancy – Supporting Language Communities Worldwide. Retrieved May 2, 2023, from https://languageconservancy.org/
″. (n.d.). ″ – Wiktionary. Retrieved May 2, 2023, from https://data.census.gov/table?q=American+Indian+and+Alaska+Native&tid=DECENNIALPL2020.P1Endangered languages: the full list | News | theguardian.com.
(2011, April 15). The Guardian. Retrieved May 2, 2023, from https://www.theguardian.com/news/datablog/2011/apr/15/language-extinct-endangered#data