Motivation and Research Question
I was initially interested in the resolution description column of the 311 data, because I had the suspicion that rather than relying on the data that comes from the users of the 311 platform, that being the New York City public, the resolution data, which indicates the action taken by the responding New York agency, would be more indicative of the efficacy of 311 overall.  From a quick glance at the data, I saw that a lot of closed complaints had resolution descriptions that indicated the problem was not really taken care of.  This led me to the research question of “How often are 311 complaints successfully or unsuccessfully addressed depending on agency and type of complaint?” 

Audience and Goal
My audience is primarily the people using 311, the New York City public.  The goal of my visualization is that a better understanding of what 311 is and is not able to do can help citizens use 311 more thoughtfully as well as highlight areas where 311 is lacking in efficacy.  While my visualizations could be used by the designers of the 311 platform to make certain improvements, I think it is more efficient to encourage users to reevaluate how they use the platform than try to redesign it.  Summarizing response details from the entire 311 dataset allows users to see general trends amongst the responses that are impossible to see from just their own submissions, and thus reshape how they use 311 for maximum efficiency. 

My Dashboard

My dashboard consists of three pie charts which illustrate how often 311 complaints are successfully or unsuccessfully addressed separated by agency.  The pie charts use three categories that I created myself to summarize the resolution descriptions: successful response, unsuccessful response, and no response possible.  I looked through every resolution description that occurs within each agency and decided myself which resolution category it fits into.  As such, the terms successful and unsuccessful are based on my personal opinion and some people might not have the same interpretations as me.   The main takeaway from the pie charts is that, for each agency, a majority of complaints are responded to unsuccessfully, meaning the agency attempted to respond, but could not take care of the reported issue for a variety of possible reasons.  For successful responses, DSNY has the best chance of success, followed by NYPD, and then DHS.

Underneath the pie charts are bar charts that explore deeper into each agency by using complaint types.  I picked the four most popular complaint types (DHS only has two complaint types) and created sectioned bar charts that show the response categories of each complaint type.  Within each bar, the individual resolution descriptions are further sectioned and can be seen by hovering over certain areas of each bar (accomplished by using complaint description as details in the bar chart).  It was important for me to include the resolution descriptions themselves, because this is how a person can make very detailed observations about the efficacy of certain complaint types.  For example, it can be seen that homeless person assistance requests are overwhelmingly unsuccessful because the homeless person is either no longer there when DHS responds, or they refuse assistance.  Similarly, residential and street/sidewalk noise complaints are overwhelmingly unsuccessful because the reported incident was not actually a violation of anything, and even successful responses at large do not require police intervention, likely meaning a simple conversation took place between the agency and the noisemakers.  On the other hand, missed collection and illegal dumping complaints to DSNY are more likely than most other complaints to be successfully responded to, because trash is definitely a violation and it does not go anywhere without someone moving it.  Derelict car complaints are almost never successful, and the resolution descriptions illustrate possible misconceptions people have about what makes a car derelict; most derelict car complaints cannot be acted on because the car has registered license plates or the car is claimed by its owner, even if it never leaves a certain spot.  

Improvements
I think there are a lot of ways this visualization can be expanded upon.  The easiest way would be to include all the agencies, although I left out the Department of Transportation specifically because its resolution descriptions redirect to their own website rather than provide useful information.  Furthermore, I think the bar charts are more useful to the public than the pie charts, because I realize now that it is unreasonable to generalize that a certain agency is better or worse at responding to 311.  The agency is really just about division of labor, and the important information in terms of utilizing 311 effectively lies more in the complaint type.  If I were to keep working on this visualization, I would delve more into complaint type to try to find more relationships between complaint type and successful or unsuccessful resolution, and find a way to highlight the more interesting conclusions.  

Link to project on Tableau Public: https://public.tableau.com/views/311Responses/311ResponsesDashboard?:language=en-US&:display_count=n&:origin=viz_share_link