Introduction
Since moving to New York last year, I have been pleasantly surprised by the amount of community-based initiatives I have seen throughout the city. Not only are these initiatives valued, common and often broadcast, but they are also consistent across the boroughs. Having lived in other large metropolitan cities, I found this stark difference in value and frequency of community-based efforts intriguing. In parallel, I have also discovered and learned about the process of Participatory Budgeting (PB) in New York and its expansion across the city. As recently as this month, District 33 (covering several neighborhoods in Brooklyn) will be voting on PB projects, and it is clear that this is increasingly a part of New York’s democratic budgeting practices. Therefore, I thought that PB would provide an interesting lens through which to assess the impact of community-based activity in NYC. By using a dataset with PB project dimensions spanning five years, I hoped to gain a better understanding of the following:
- What types/categories of PB projects were most popular at the local level;
- How project funding priorities differ between communities;
- Whether certain geographic areas engaged in PB practices more extensively or regularly or devoted more financial resources towards it;
- Individual and general costs associated with PB projects and implementation.
Visualizations that Informed my Design
I found that there were not a lot of visualizations available for PB data; the ones I did find were more helpful in providing possible directions to explore with the data. They did not offer effective visual guides for how to display PB data and were a good indication of what not to do.
Figures 1-3 (from the New York City Council website) detail the results of cycle 5 and 6 PB voting across participating districts. I found the concept of mapping projects to distinct geographic locations helpful as it can show us trends and project volumes across the city.
Figure 2 was interesting as it provided a breakdown of PB projects per category. However, the visualization was also a great example of how not to show categorizations. The chart format makes it hard to see differences between slices with similar values, and there does not seem to be any logic to the color choices and what they are meant to represent. These shortcomings are also apparent in Figure 3, where the colors are again meaningless and do not even relate to the category colors. Nevertheless, having costs associated with a district or category would be helpful to identify further differences between the projects and where they were implemented.
Materials
To produce this visualization, I used Tableau Public. Each chart/graph was created individually and then collated in a summary dashboard. The data for this visualization was produced by the City Council of NYC and was found on the NYC Open Data website here. I then used OpenRefine to clean and sort the data before moving it to Tableau Public.
Methodology
I chose this PB dataset for two particular reasons:
- It spanned five years (2012-2017), allowing me to better illustrate and identify PB trends, increase and movement time;
- It contained several different qualitative and quantitative dimensions that would provide interesting insights both individually and when displayed together and filtered as a group.
Once the data was finalized in OpenRefine, I imported it into Tableau to begin building my charts. Each one was created individually and compiled in a dashboard. Having the charts together would allow side-by-side comparisons, but it would also allow filters to be applied across charts, so a specific dimension across different PB project categories could be assessed. I wanted to create charts that were different from each other and contained specific information, but that also had at least similar or exact shared dimensions to allow for cross-chart filtering and assessment.
Data clarifications:
- About 45% of the data did not have longitude, latitude and/or NTA values. As the dataset was comprised of 1,491 rows, even after these were removed, I felt that there was a significant enough number of projects displayed to allow us to illustrate trends and differences between PB projects.
Results and Interpretation
The results indicated that there were specific PB project categories that benefitted disproportionally from a higher volume of projects. Education received 2.5x more projects than the next highest category, Schools. I would be interested to see whether this is due to a genuine need or whether any other biases or limitations are affecting the number of PB projects submitted and approved for different categories.
The map was encouraging as it seemed that the amount of PB projects, even the more expensive ones, were well-distributed throughout the boroughs. Other results were more disappointing such as identifying that Public Housing only received projects in 2013 and even then, it was only 7.35% of that year’s PB projects. Lastly, the overall yearly costs confirmed my assumption that PB as a practice was becoming more common across the city.
Reflections and Future Directions
It was my second time working with OpenRefine and Tableau, and I felt much more comfortable with both tools. This allowed me to be a bit more exploratory in the charts I created and think I have achieved more varied visualizations as a result. On the other hand, due to limitations with the dataset and Carto, I was not able to engage as deeply with the geographic level data as I would have liked, particularly on a district by district level.
In terms of future directions, I think it would be fascinating to work with PB data from other cities (both domestic and international) to have as a comparison to NYC. This would allow me to validate my initial assumption, that New York may have a more significant and varied spread of community-based initiatives, particularly when it comes to PB. It would also be interesting to understand whether New York is unique in encouraging adoption of democratic budgeting processes and also how PB projects vary across countries and regions.