As someone who has been focusing on data analysis for the past years, professionally and academically, I always find myself re-centering around the lens of ethical data collection and use. Simply because a data set is interesting or may hold new findings does not mean that said data ethical or that you are the correct individual to be using said data.
This was a strong undertone of Jennifer Guiliano and Caroline Heitman’s article, “Difficult Heritage and the Complexities of Indigenous Data”. As scholarship and the empirical space embrace the importance and historical neglect of indigenous populations after centuries of colonialism and erasure, it is important to keep in mind that when trying to research indigenous histories, the line between highlighting colonialism and perpetuating it is incredibly thin. Guiliano and Heitman did an excellent job of displaying how these practices are continued today in the, maybe well intended, but absolutely harmful, non-indigenous preservation efforts. While the violence, genocide, disease-spreading, and overall abhorrent practices of colonialism have done irreparable harm, so have these ill-advised 20th century preservation efforts.
The digitization of indigenous data without mindful consideration of how this preserves harmful practices lays the foundations for continuous harm. When this data is collected, manipulated, and digitized by non-indigenous peoples for their own use, the perspective and usage of such data continues to be centered around colonialism.
Interestingly, this creates a worthy intersection with some of the points made by Lev Manovich in “What is Visualization?”. It is in this article that Manovich lays out how data reduction methods are not only outdated given our modern digitization technologies, reduction is somewhat a means of erasure for the complexities of large data sets. It is in this unneeded erasure that nuance and value are lost from meaningful data.
The intersection of these articles makes a clear point regarding data practices, namely in the digitization and use of Indigenous data, but perhaps also creates a lens that should be applied to all data practices. Not only should researchers be questioning “should I research this?”, but also need to question “am I the right person to be researching this?”. Digital humanists hold a lot of power in the presentation of their findings, and with that comes the ethical requirement to question whether their research and practices being used to complete it is undoing harm or preserving it.
Census forms may be the place where America’s bizarre compromises related to race and citizenship appear most clearly. I’ve been exploring two datasets related to the Census for possible projects. One is the 1865 Census taken by New York State. The other includes a tabulation of that Census and election-district level returns for three failed constitutional amendments extending suffrage for Black men. Two researchers compiled this data in 1984 from the Census books and New York State Assembly records, and it’s now available in the Inter-university Consortium for Political and Social Research repository. I also have a third source, tabulations of the 1865 Census entries published by the New York Secretary of State in 1867. Please note that I am not sure how to handle the demeaning language used in the Census. I’m repeating it here to discuss a harmful categorization process but I need help to know if I’m doing that in a way that is respectful. Also please note that the nationalities assigned to White people are random. I wanted to see what I could do with nationality data in a visualization but matching them accurately for White people is a big project that I haven’t completed. The total counts of different countries of origin are accurate but I don’t know the gender ascribed to them in the Census and I don’t know how many were over the age of 21. I could do it for Black residents because there were so few of them and for Tuscarora residents for very questionable reasons discussed below.
Voter Categories
New York State held referendums in 1846, 1860, and 1869 asking if a property requirement for voting, which did not apply to white men, should be eliminated. These amendments were about moving some people into a new category: “native voters.” Voters have two columns in the 1865 Census: native or naturalized.
White men over the age of 21 born in the United States with one year of residency in New York were native voters.
There are two categories of non-voters: “aliens”, and “colored person not taxed.” All of the immigrants in the Census book are described as white. They could gain citizenship by renouncing any hereditary titles or loyalty to foreign governments, and vote without property restrictions (but they will always be marked in the Census as a different category than native voters).
Men described in the Census as Black or mixed race go in the native voter column if they have lived in the state for three years, own property worth at least $250, and pay taxes on it. Very few people met the $250 threshold regardless of their other categorizations.
All of the referendums failed. The majority of Black men could not vote in New York State until the federal 15th Amendment was ratified in 1870.
Lewiston, New York
I made a typical dashboard using the voter returns and Census books for one town in Niagara County. I chose Lewiston because it’s small enough to work with (the Census has 2,998 entries) and two Black voters lived there. The other places with Black voters had too many people for what I wanted to do.
Lewiston is on the Niagara River (which is the Canadian border) and just north of Niagara Falls. It completely surrounds the Tuscarora Reservation (the Library of Congress has a map of Lewiston in 1860 accessible here). The Niagara Falls Underground Railroad Heritage Center has compiled evidence of significant activity in Lewiston and throughout Niagara County, including several safe houses and residents who had been enslaved and either settled in Niagara County or travelled through on their way to Canada. There was also significant organizing for abolition and before the late 1860s, organizing for universal suffrage for men and women. It is too easy to incorporate this history into myths about “free” states. The referendum votes add cracks to overly simplified stories about Northern states like New York.
Lewiston residents voted against the referendum, as did Niagara County overall:
I can complicate the yes or no story by showing how close the vote was, how many eligible voters did not participate, and how many were not allowed to. I can also nod towards intersectionality by showing gender, race, and nationality in one table. But using these sources and presenting them in this way requires adopting their “cultural logic” as Cottom describes.
Implied Precision and Stability
The cultural logic of Census forms is that race, gender, and nationality are distinct, factual categories and it is possible to count people according to those characteristics. Visualizing Census data in bar charts and tables requires obscuring the reality of Census data, which is that it is a mess. When you look at original hand-written Census records, you are more aware of the process. Things are misspelled or scribbled out, the handwriting can be hard to read. None of the population totals match across my three sources. Further, someone looking at one Census book might not understand how much movement occurs between categories. Even place of birth frequently changes across Census records for the same people.
Implied “Sameness”
Drucker’s visualizations show subjective experiences of time and space related to the point of view of the person experiencing them. This vote meant different things to people based on their bureaucratic identities, their personalities, and their life experiences. The first failed vote was experienced differently than the second, and the third. Showing that many people did not vote, even as others fought for the right to do so, points to different subjective experiences but it doesn’t visualize them. A person whose right to vote was in question probably experienced this event very differently from a person given the authority to decide that question for someone else.
Lack of Situated Knowledge
The state legislature tabulation for Lewiston includes a footnote: “in addition to the above 572 Tuscarora Indians are reported as residing in the town of Lewiston.” Population tables about the Tuscarora and people living in other reservations are in a separate section.
Everything about how I treated the people living in the Tuscarora reservation is questionable. Guiliano and Heitman describe how data has been used to fragment Native American and indigenous communities, and asks scholars to work with those communities before using data collected for the purposes of colonization. They describe a different data culture that necessitates collaboration. There are so many questions I can’t answer without doing this. For one, the Tuscarora Nation is not “in” Lewiston. I don’t know how someone living on the reservation would feel about being included in a Census of Lewiston. I assigned everyone in the reservation “Tuscarora” as a nationality and “American Indian” as a race, which is what the category is in a modern Census. But I doubt everyone living on the Tuscarora reservation are considered Tuscarora members. Even if they were, the word nationality may be inappropriate, and it may not make sense to treat it equivalently to a term like “American” or “German.” The population tables do not include race, so I can’t show that people of other races lived there. But I know that they did from other sources.
It felt wrong to ignore the Tuscarora because as Guiliano and Heitman describe, I’m in a data culture that values more and more public data. But the Census is such a harmful document that it is wrong to carelessly use it. If this was a real publication I would treat that data differently.
Strategy: Combining Close Reading with Distant Reading
D’Ignazio and Klein in Data Feminism and Cottom in “More Scale, More Questions,” describe restoring context and situated-ness to data visualizations as a strategy. I tried doing this in three ways: I re-created everyone in the Census as colorful shapes, added annotations, and added randomized numbers to avoid sorting people so strictly by category.
This is everyone with a Census entry in Lewiston in 1865:
In some ways this lessens the false preciseness because you can’t read it as a specific tally of any group. It might help a reader see that there are many possible points of view of the referendum vote, but not much more than the bar chart or table above. What I really wanted was a way to make the categories appear more porous. It’s still really sharp shapes when they should be fuzzy.
For this one, I added random numbers which are the coordinates for each line. I also added annotations with information about specific people or information that is relevant to an individual person’s experience:
I like that this is hard to read from a humanities standpoint even though it isn’t helpful for understanding the data. It reminds me that I’m never really seeing the people themselves when I’m looking at a Census entry. When I collect Census entries on people it really feels like I’m “seeing” them, but I’m not. It reminds me that data is taken.
It fails in terms of demonstrating how different people would have different experiences of the event. The people whose citizenship is in question are nearly impossible to find. In Tableau, you can pull them out with filters, which helps a little bit. The annotations may also help.
For the praxis visualization assignment, I explored a rock music list a friend gave me. It was shared as an excel file, so exporting as a CSV for Tableau and Palladio was made simple. The list includes music artists, specific song(s) from that artist, the genre(s) for the song, and the year the song was released.
Palladio
After loading the data in, I could experiment with the map, graph, table, and gallery views. Map view wasn’t useful as I do not have location data in this data set and table view mirrors what I have already done with this list in Notion. The graph view was the most useful for seeing the data in a different light. I used artist as the source and genre, song, and year as the targets. Each time, the changes in relations between the information surprised me. The first time I used this list, I just started listening based on the order in the excel file. With the graph view, I was experiencing the data as a journey; listening to a particular song by an artist and then another because of the relation they share by genre or year released. Overall, this was the most appealing visualization between the two platforms I applied my dataset to.
Tableau
Previously, I tried this dataset out last week by attempting to add country tags for each artist and create a map with it. I found that the list skewed heavily towards the UK and the USA. I didn’t continue down this route for that reason and there are like 300 or 400 entries that would need to have a location added. Even though it didn’t work out with making a map, I wanted to try the dataset again to see if I could still get some insight (and get more experience with Tableau). I watched some tutorials to better understand how adding dimensions and measures would affect a visualization.
I made three worksheets. The first is a look at the artist by year. This view didn’t tell me much and because there was an abundance of artists included, it just felt like an endless scroll. The second worksheet was a look at entry count by year. This highlighted a heavy concentration in the 1960-1980s. The third view was the frequency that an artist appeared in the list overall. I think the last two views were the most illuminating.
Using the dataset in Tableau was a more outwardly analytical experience. I am looking forward to discussing the list with my friend. Some questions I have are: what was his methodology for compiling this list, and did he realize that there was a favoring of Western countries? I was wondering about the term “Intro” in this context. He named the list “Intro to Rock” and I wonder if his interpretation of intro means that it should start at what we believe to be the origins in time for rock. “Difficult Heritage and the Complexities of Indigenous Data” and “A Review of ‘Two Plantations” were most impactful in giving me examples of colonial impact in data and data representation, and what it could look to collect and present data more ethically. As I was working on this, I noticed more immediately than I would have previously that the list is more male centric and favors Western artists. As a list between friends I don’t feel it’s a huge issue, but I am reminded of the numerous times I’ve encountered recommendation list for popular media. The lists often reflect this same issue of representation, where cultural “best of“ is concerned.
Tableau is a great tool for visualization, although there are some learning curves. Charts like the ones I have below are on a high level intuitive to create because you can drag and drop your data as well as the x and y axis classifications and the information will appear before your eyes. Then you can choose between a myriad of plots to best represent your objective. Below I have total checkouts of the San Francisco Public Library by age classification and the total by each home branch. I wanted to compare these because of the large uptick of checkouts in 2015, I was searching for something in the dataset that might explain why so many more books were being checked out in just a year. For this assumption I hypothesized that it was a localized uptick. While the Main branch in pink did have the largest uptick overall, all the other branches seem to trend in the same way where there is exponential growth between 2015 and 2016.
In further exploration I decided to look at the demographics of the patrons checking out books. Age was the only identifier of the dataset and as seen above, adults were the most represented classification. This is not surprising, but I thought it would still be helpful to see the amount of people classified as adults by comparing the age ranges used. Juvenile, young adult, and seniors, represent the margins of the age ranges and therefore there are less of them overall. Adulthood covering the ages of 18-65, make it the most encompassing classification. The accompanying line graphs show the trend of adults checking out books at the Main Library and all libraries excluding the Main Library. In an effort to reinforce that these are only showing the adults in each age range I have made the key a gradient of blues to match the key for adults in the bar chart.
In order to more effectively show only the data that pertains to my purposes I used the filters in tableau. You are easily able to filter each unique value from a column. For the line graphs throughout I have filtered out the years before 2009 because there was steady rate of checkouts that showed little difference year to year. The most interesting points in this data happen from 2012 onward, but I did not want to completely get rid of all the years before that because the growth in 2012 is more significant when compared to the steady state of checkouts from the years before. Other filters used were Age Ranges and Patron Type Definition.
The final assumption I made in attempt to figure out why there was such a spike in library checkouts was increased membership. In the charts below I have shown the breakdown of patronage sign up by year since 2003, I am using the pie chart to act as a key for the accompanying line graphs in order to reduce redundant information, there could be a better way of doing this, but with limited tableau knowledge I have fit these charts together in premade containers tableau provides.
The most surprising part of looking at this data was finding out that the uptick in checkouts were due most heavily by patrons who signed up in 2003. This made me pause and wonder how accurate this part of the dataset was. We can see from the pie chart that the amount of patrons from 2003 is large, but similar to recent years of 2015 and 2014, and we don’t see the same jump in usage. This ultimately leaves me with more questions about the data overall and no real hypothesis for why the number of checkouts has changed so dramatically.
I enjoyed using tableau as a way to easily visualize data and quickly check my assumptions. This is not practical data analysis as a whole, but an interesting way to quickly see what you have without need for other specific software knowledge. And with increased knowledge of how to use the platform more complex images can be produced. But with basic information and intuitive design, you can easily display important factors from your data. There are some drawbacks though, not all datasets will work well with tableau’s platform and found that there is a particular way datasets need to be organized for tableau to read the information easily.
I’ve become increasingly aware of citations as a research tool to dig deeper into a topic and the importance of attributing credit to scholars for their work. Of particular interest to me, are the links and citations in Lev Manovich’s “What is Visualization?” and Tressie McMillan Cottom’s “More Scale, More Questions: Observations from Sociology.” In both articles, we are exposed to visualization methods in “cultural texts, […] poems, paintings, music compositions, architecture, or, more recently, computer games, generative artworks, and interactive environments.”
In Manovich’s piece for example, the advent of information visualization, has allowed us to observe historical events that are expressed artistically, to exhibit “many moments spread across time and bringing all of them together in one single moment to create something new.” A timeline of events in history are cleverly displayed to remind us of our country’s economic struggles and advances, its cultural trends, its tragedies, and successes. The progression of these images in this timeline of magazine covers also reflect advances in technological developments, such as with the use of drawings, paintings, photography, “contemporary software-based visual(s),” via color, clarity, and hue; namely from applications that were once only available in black and white.
Although “synecdoche” is referred to in Brendan Dawes’ Cinema Redux example, I feel that in the Manovich Mapping Time exhibit, also remind us that many Time Magazine covers are meant to associate something larger than just what we see in an image. In my view, these associations are often taken for granted, including the time it may have taken for both the original work to arrive at its final “edition” and the effort of the digital humanist(s) in reimagining this work. This exercise has caused me to be more “mindful” of the creators’ goal(s) of achieving reactions of awe from those of us (or at a minimum me) that study or observe their work.
This brings me back to the relevance of citations of the written word, but with a somewhat different significance. Here I’m more interested in McMillan Cottom’s argument on how:
“Sociology has developed a diverse toolkit to identify, measure, and analyze various forms of text with an attention to political economy. This includes content analysis (e.g., newspaper content), organizational analysis (e.g., texts produced by institutions or organizations), and quantitative narrative analysis or QNA (e.g., a sociological complement to distant reading).”
She cites: Franzosi, Roberto, Gianluca De Fazio, and Stefania Vicari. “Ways of Measuring Agency: An Application of Quantitative Narrative Analysis to Lynchings in Georgia (1875–1930).” Sociological Methodology 42, no. 1 (2012): 1–42. Where the structure of sentences and the use of grammar has provided an understanding of literature by aggregating and analyzing data. Rather than reading things “up close,” the authors of this article insist that by reading a book (i.e., a Victorian novel), one “can’t uncover the true scope and nature of literature…because the sample size is too small.” Furthermore, the grammar of a particular era, as written in literature, newspapers, police logs, etc., have properties or nuances that the reader can’t detect without the aid of computers and coding, which “provide graphical representations of the relationships between social actors taking advantage of… the subject-verb-object (“SVO”)” and their modifiers, to produce a powerful “story grammar.” Here’s an example of the sourcing of “massive amounts of data” to produce one of the maps of the lynching atrocities in Georgia between 1875 and 1930:
I have some experience with data visualization tools like Looker Studio by Google and Power BI by Microsoft. Tableau caught my attention as a tool that could offer valuable insights. Initially, I intended to work with birthday data and correlate Zodiac Signs with participation, but I found that managing this data was quite challenging. Consequently, I opted for an alternative approach – analyzing my clothing items to distinguish between thrifted and fast fashion pieces.
I focused solely on my everyday wear, which includes shirts, pants, jackets, sweaters, and T-shirts. I categorized each item into one of four categories: Fast Fashion, Thrift Shop, Hand Me Down, or Small Business. Additionally, I assigned secondary attributes to the clothing items. To facilitate this analysis, I created a spreadsheet and conducted a qualitative assessment of my wardrobe, which I later imported into Tableau for visualization.
While working in Tableau, one of the challenges I encountered was ensuring that the graphs had consistent colors to establish connections between them. Additionally, some of the terminology within Tableau made it a bit difficult to navigate the platform.
Nonetheless, I persevered and experimented with various charts and graphs, ultimately selecting the Bubble Graph for a comprehensive closet breakdown. This project served as an exploratory data analysis, as I had preconceived assumptions about my wardrobe composition. However, I sought to validate these assumptions and uncover any intriguing findings that could influence my future shopping habits.
In summary, my exploration in Tableau was not only aimed at confirming my initial beliefs about my closet but also at discovering new insights that may shape my future fashion choices.
Here is the visualization of My Closet Breakdown from Tableau.
For the visualization assignment, I chose to work with NYC Citibike data. The data (so far) is presented as 123 separate csv files spanning 10 years. Some interesting observations while processing the data:
From June 2013 – January 2021, Citi Bike data is stored in the following format:
'tripduration','starttime','stoptime','start station id','start station name','start station latitude','start station longitude','end station id','end station name','end station latitude','end station longitude','bikeid','usertype','birth year','gender'
From February 2021 – July 2023, Citi Bike data is stored in the following format:
Post February 2021, gender, bikeid, and birthyear are no longer stored. Instead rideable_type and ride_id are introduced. The former stores the type of bicycle (electric, classic, docked) while the latter is a unique hash sequence. Unlike bikeid, which can be tracked across records, ride_id is a blackbox that is only useful as an index.
I used Plotly to create the visualizations and MapBox for the maps.
While some of the visualizations were interesting, the way the data was collected, structured, and presented limited what could be visualized.
Data then is an expression of those systems and worldviews and thus must be recognized as intrinsically tied to how the community defines itself.
Cultural Analytics Jennifer Guiliano and Carolyn Heitman
The data captured is more of a reflection of preferences and priorities of the Citibike system than a true representation of experience. Many of the factors we would consider part of the experience of using a Citibike (let alone biking in NYC) were unaccounted for.
Throughout the process of creating these visualizations, there were multiple levels of abstraction in play. The sheer volume of data, the structure of the dataset, and the resulting visualizations. Each layer was another step removed from experience.
In an effort to make the data more personal, I used the bikeid datapoint to track starting and ending stations over the course of a year. However, much like the bike itself, the data (and visualizations) are tied to the system within which it operates.
I’m not very experienced with coding / programming / quantitative data analysis and I’ve been working with mostly qualitative data and so I was looking into what my options are for beginner-friendly data visualization that apply to qualitative data. I saw in Tooling Up For Digital Humanities: Data Visualization that word clouds might be a good fit for a data set I’ve been working with given the stage I’m currently at in analyzing that data. I tried making a word cloud in python but I haven’t tried coding in python in years and I couldn’t successful install the package that was needed for it to work. I was using a notebook in my Anaconda cloud account to avoid having to download the Anaconda navigator onto my computer but I might try it again after I install it. But honestly there’s a lot of word cloud generator tools out there that seem pretty simple to use so it might not be worth my time trying to create one in python. I was able to easily copy and paste my data into the generator on https://www.freewordcloudgenerator.com and it gave me this. I feel like it cut off some of the words though and I’m not sure how to fix that so I might try another generator.
I also tried analyzing the same data set (they’re TikTok comments) on Palladio. I couldn’t do much since it seems to be a tool designed for quantitative data visualization or at least I couldn’t access some of the features that I might be able to use later once I have the results of my qualitative data analyses but I was able to just use the one quantitative variable (number of likes each comment got) to create something in it. I liked how it lets me visualize all of the comments in one place and move them around and set the node sizes to be bigger for the comments that got the most likes.
I think once I finish some of the qualitative coding I plan on doing on these TikTok comments I can see data visualization tools being really helpful for helping me see the patterns in the data, since “the goal of information visualization is to discover the structure of a (typically large) data set. This structure is not known a priori; a visualization is successful if it reveals this structure” (Manovich, 2010). I’m also really interested in how data visualization tools can help me with the task of demonstrating how epistemological violence shows up in these comments. A Review of “Two Plantations” mentions how the discussed data visualization project, “produces a conversation within the site or among users visiting about who and what is missing… It is both what is present in the history and what is not that resonates with users”. I also need to figure out how to show what isn’t there, how to provide the necessary historical context to show the gaps in the discourse I am analyzing and showing the systemic reasons why the “more reliable sources” people kept asking to provide them in these comments don’t exist.
Many of this week’s readings resonated with my experience and concerns working with and adjacent to “tech” over the past few years. I actually decided on CUNY’s quantitative methods masters program specifically because it was grounded in the social sciences and intentionally offered the flexibility to combine quantitative courses with more theory-driven learning. I’d become increasingly concerned with the lack of inquiry and critique being applied to data usage and other quantitative processes.
For this reason, I especially appreciated Johanna Drucker’s argument that everything we know as data should actually be called “capta” to more accurately reflect the ways in which we “take” information, as opposed to “data” indicating that this information is given for our presumably neutral observation and analysis (Humanities Approaches to Graphic Display, 2011). As Jennifer Guiliano and Carolyn Heitman detail, “data in the humanities are always subject to the systems of knowledge that were used to capture, represent, and disseminate them” (Difficult Heritage and the Complexities of Indigenous Data, 2019). I think a humanist and/or ethical lense is exactly what’s missing from so many of the conversations around the today’s technology and that of the future. I also recognize that this is not a new perspective or a new challenge for systems of power, as indigenous peoples and other systemically oppressed populations can attest to. The danger of tech and data, as all the articles explain, is that they are often presumed to be neutral and even helpful tools instead of agents of surveillance and exploitation.
As recent efforts have exposed, there’s always been a lack of transparency regarding how much of the technology we use works. My own journey learning some coding languages has revealed to me how accessible the underpinnings of technology are, when the time is taken to explain and explore them. For this reason, I especially liked Dr. Tressie McMillan Cottom’s explanation of her process for analyzing a seemingly simple question through text analysis as she illuminated a number of junctures where intentional decisions and/or assumptions were made in order to facilitate an analysis (More Scale, More Questions: Observations from Sociology, 2016). This level of transparency and intention applied to things we use every day like Google’s search engine or Google maps or an Apple phone might create a more expansive and accessible world of technology for all its users. This leaves me asking questions like: where are there more examples of what this could look like in common practice? What leverage do we have beyond advocacy and organization to make some of these changes happen?
I used Tableau to create a visualization comparing Craigslist furniture listings in the three largest cities in the United States: New York, Los Angeles, and Chicago.
Craigslist is fairly easy to scrape because the format is consistent across pages. I used the BeautifulSoup and pandas packages in Python to pull the listing title, price, and listing location (i.e., neighborhood, town, borough, zip code) of furniture listed in for sale in each city into a dataframe, which I exported as a csv file. I did not “clean” the listing titles or locations, and I’m glad I didn’t, because retaining those two fields in their “raw” state raised some of the questions with data visualization that Johanna Drucker and Lev Manovich discussed in their respective articles.
Once I had the data, I went to Tableau. The public platform is robust and allows users to upload tabular or text data for analysis and visualization, and to combine those visualizations into a dashboard. Users can upload more than one file and combine individual tables as you would with relational databases, but you don’t need to know SQL in order to do so.
Initially, I found myself struggling to actualize graphical representations I had in my head as a result of the crowded interface and levels of abstraction. I am used to more explicitly transforming data with functions or formulas through code, with a more minimal visualization functionality in Periscope/Sisense, which is a different business intelligence platform that connects directly into your database and generates charts off code. It took time to figure out where I needed to drag-and-drop a field in order to add it to the x vs the y axis. For example, the first time I dragged the City field into the workspace, it generated coordinates and a map popped up. Its ability to infer what format to present your data in is both helpful and frustrating.
The platform also felt grounded in traditional info visualization the favored bar charts, pie charts, and line graphs, regardless of the data uploaded, which makes sense given that it is largely an enterprise software. I’d also wager this is why Tableau offers little documentation, favoring few example videos and recommending user join a community of others to learn.
With that being said, I was able to explore the dataset and create a final dashboard that made use of statistical charts, while also finding the edges of it’s utility in my current exploration, with my current knowledge of the tool.
In trying to reconcile Neighborhoods, Boroughs, and towns outside from each city, I was reminded of what Johanna Drucker said in “Humanities Approaches to Graphical Display”: that no data pre-exists its parameterization. They are already interpreted values. Any patterns observed or conclusions drawn from location with such fluid boundaries would be subject to criticism given the contested history of neighborhoods in major US cities. In addition, this data was user generated, which might make it easier to understand that the labels and classifications are subjective and influenced by the person creating the listing. I wondered about ways I could use Tableau in ways that were more temporal or flexible, as Drucker described.
I also found myself wanting to explore direct visualization, since I took interest in the listing metadata. Listing titles were long strings of keywords, and the locations were far from standard. I attempted to combine similar or duplicate locations that followed a slightly different format, but debated whether or not that was my role, for reasons mentioned above. The uncleaned listing titles, locations, and images feel critical to understanding cultural, social, and economic ideas underpinning craigslist, yet it was difficult to find ways to incorporate the media objects themselves, other than adding in screenshots and returning the fields themselves.
Ultimately, the platform does enable laymen to create well-crafted visualizations, and I am please with the initial visualizations I put together.
Need help with the Commons?
Email us at [email protected] so we can respond to your questions and requests. Please email from your CUNY email address if possible. Or visit our help site for more information:
