This lab assignment was both daunting and exciting. I had an idea to look for data concerning New York City, though that was open to numerous subject areas – I originally envisioned utilizing maps of the boroughs somehow, though this did not pan out given my chosen dataset. By deciding to work with data concerning elected official donations to not-for-profit organizations in compliance Chapter 9 of Title 3 of the New York City Administrative Code and the Rules of the Conflicts of Interest Board, I opened up to working on a more national scale with an emphasis on donation values and organizations involved in the process. With both its expansions and its limitations, the project proved rich with opportunities to display and share information about New York political donation records.
I perused a few dataset sites before finding the one I ended up utilizing for the project – these included Data is Plural, Awesome Public Datasets, NYC Open Data, and Data.gov, the lattermost being where I came across Donations to Not-for-Profit Organizations Affiliated with Elected Officials. I wasn’t intentionally looking for any specific topic with my dataset; my main search strategy was looking under New York terms in hopes of finding something interesting and relevant to the city I live in. Donations from elected political offices seemed like a straightforward area that would be both illuminating to interact with while providing the possibility for visualizations on multiple axes.
In order to create these visualizations with my data, I used the Tableau Public platform. This application allows users to play and explore with the variables of their dataset in order to create different forms of informational visualizations.
I started to work with my dataset by running it through OpenRefine, an open source application that helps clean data. The one major issue with the data I chose to work with was that of incorrect data entry, which included spelling Bill de Blasio at least five different ways, with some including the title of Mayor and some not. I used the text facet and cluster tool to group all of these data points together to refer to the singular candidate, which was tremendously effective in making my visualizations clearer, more succinct, and simply more accurate. I then input this cleaned dataset into my visualization software, Tableau Public, which greatly helps beginners orient to its interface by suggesting different kinds of maps, graphs, and tables that one can use given the kinds and number of variables that you add to the workspace. I knew I wanted my visualizations to stylistically look different from one another, so this involved much work in experimentation to see what different variable combinations could produce.
As I tried out a variety of these visualizations and combination possibilities, I had my girlfriend provide feedback on what did and did not work. Originally, my second visualization (“Donor Amount by State of Residence (2018-2020) (Excluding New York)”) was a bullet graph that listed states on the Y-axis with a dot for each year at least one donation hailed from said state. She pointed out that this depiction needed much more context to truly understand what it was saying or why it mattered, and the singular dot used to represent an unidentified donation amount or amounts was difficult to comprehend and lacked any comparative element between the states. I ultimately switched to the tree map to better show the make up of donations state by state, excluding the overwhelming majority coming from within New York, through easily compared sizes and shades of rectangles.
Additionally, in order to maintain a level of diversity in my visualizations, I also changed an early version of candidates by donation amount via a packed bubbles format into something else entirely – a pie chart representing the value of in-kind donations by their donor name (“Value of In-Kind Donations by Donor (2018-2020”). This was in response to her critique of the original work, which again lacked context; one large bubble represented the donations made to Bill de Blasio from 2018 to 2020, while Eric Adams’ bubble as Brooklyn Borough President was much smaller. This makes sense, given their offices at the time, but even with the years provided, users could be confused as to why the data is so disparate or why it does not represent officials across a broader spectrum. While this is the data itself falling short and not my visual interpretation of it, I decided to scratch that version entirely to instead try and represent something that better serves my assignment’s goal of providing clear, quality visualizations.
The other two visualizations, “Officials by Donor, Donor Amount, and Year” and “Value of In-Kind Donations Vs Donation Amounts” came about to their final product with less issue and external critique. The former depicts a table at the intersection of elected officials, donor organizations, subsequent donor amount, and the year of donation. Given the higher number of variables, I see it as one of the most contextualized visualizations while remaining relatively simple to understand. The donation dots are legible, while including the date shares information on how certain donor amounts from the same organization may have shifted over the relevant years.
The latter is a line graph showing the difference between in-kind value donations and monetary donation amounts over the three years of data collection, with donations climbing continuously while in-kind values drop dramatically in 2019. The two lines are easy to follow and interpret while portraying trends that would be intriguing to research further.
The dataset itself had a lot of blatant limitations by only including donations made to de Blasio, Adams, and a few for Steven Matteo, as well as a few for the unnamed office of the Mayor, which I left in case there are such donations made through the city without being explicitly affiliated with the official themselves. Naturally, as mayor during the years the data are pulled from, 2018-2020, de Blasio’s numbers are overwhelming in comparison to the other candidates, making direct comparisons both unhelpful visually and as conveyors of information – it isn’t news that the mayor incumbent provides more donations than a borough president in the same year. This forced me to approach the data in less obvious ways, which I appreciated in my exploration of what was possible in Tableau. Going forward, I would be interested in looking at data that draws a comparison between different terms of the same office – how do Eric Adams’ donations as mayor today compare with de Blasio’s, both in early years and latter?
NYC OpenData (2020-2022). Donations to Not-for-Profit Organizations Affiliated with Elected Officials. [Data set]. NYC JSON. https://catalog.data.gov/dataset/donations-to-not-for-profit-organizations-affiliated-with-elected-officials