Introduction

The older adult population of the United States is projected to grow from 46 million to 98 million by the year 2060 (APA, 2019).  Moreover, the American Psychological Association Office on Aging (2019) reports that “one out of every four 65 year olds will live past age 90, and one out of every 10 will live past age 95” (APA, 2019).  Consequently, the number of Americans who are 85 and older is currently the fastest growing age group in the country (APA, 2019).  Thus, researchers, health care providers, and policy makers have begun to focus their efforts on how to ensure healthy aging and improve quality of life for older adults (APA, 2019).  

The City of New York is no exception when considering the expansion of the older adult population, and reports that approximately 1.64 million older adults currently reside in New York City and that by 2040 the older adult population could reach 1.86 million (20.6% increase; City of New York, 2019).  As a result, the New York City Department for the Aging (2019) has made it their mission to “eliminate ageism and ensure the dignity and quality of life of diverse older adults.”  Toward that end, they have partnered with numerous community-based organizations and support caregivers through “service, advocacy, and education” (City of New York, 2019).  Additionally, in 2017 The New York City Department for the Aging launched, Age Friendly NYC: New Commitments for a City for All Ages, which included approximately 90 city initiatives hoping to address the unique concerns affecting the New York City older adult community.

However, implementing citywide initiatives can be a daunting process and highlights the need for useful and relevant data sources.  When considering the City of New York, one data source that could potentially be utilized to begin addressing the needs of older adults could be the New York City 311 (NYC 311) service-request program (City of New York, 2019).  The NYC 311 program is designed to allow New York City residents to report problems and ask questions through the use of a non-emergency phone number and website (City of New York, 2019).  Following the implementation of the NYC 311 program, the service request data was made available through the New York City Open Data Portal (City of New York, 2017).  The data sets are maintained by the City of New York, are updated daily, and allows individuals to access the NYC 311 service request data in hopes of benefitting each New Yorker and their respective communities (City of New York, 2017). 

Current Project:  Data Visualization

In response to the importance of being able to make evidence based decisions for the older adult residents of New York City, the current data visualization project hopes to explore whether or not the NYC 311 service request data can assist with providing better services through a visualization of “elder abuse” complaints.  The specific research questions hoping to be addressed are: 1) Which New York City zip code reported the most elder abuse complaints? 2) What month had the most elder abuse complaints organized by New York City zip code?

Data

The current visualization project utilized the NYC 311 Service Requests from 2010 to Present dataset that can be accessed through the New York City Open Data website (City of New York, 2019).  For the purposes of the current project, the data was filtered using two different conditions.  The first condition filtered the dataset by complaints reported after the date of December 31, 2016 at 12:00 A.M.  This was chosen in order to limit the overall size of the data set, while also ensuring a useful range for the purpose of the data visualizations.  The second condition filtered the dataset by complaint type and included any complaint that contained the terms “elder abuse.”   The use of these filters produced 774 unique “elder abuse” complaints from January 1, 2017 to the present date.  It should be noted, that the NYC 311 Service Request Data Dictionary (code book) did not provide any specific definition of “elder abuse.”  Additionally, when filtered by “elder abuse” the unique complaints did not include specific latitude and longitude information.   All data visualizations were created with the software program Tableau Desktop 2019.2.

Data Visualizations

The first set of data visualizations represent the total number of elder abuse complaints made by zip code.  The use of a bar chart was chosen because it provides a simple and straight forward visualization of which zip code reported the most elder abuse complaints.  However, while the bar chart might be useful in displaying the total number of complaints by zip code, individuals unfamiliar with the specific locations of those zip codes might struggle with interpreting what part of New York City is being represented.  Thus, the visualization directly below the bar chart is a map based on the zip codes included in the current data set.  Each different color represents a different zip code and hovering over each area will display the total number of elder abuse complaints.  While both visualizations display the same data content (zip code and total number of complaints), utilizing a colorized map introduces another dimension to the visualization and further clarifies which areas of New York City were reporting the most elder abuse complaints through NYC 311.  Accordingly, the use of the map allows the user to quickly see that the vast majority of the complaints were made in zip codes representing the Brooklyn and Queens areas of New York City. This is also an example of the data informing the visualization process. It was the original intent of the project author to display elder abuse complaints by borough. However, after reviewing the data, it was discovered that Brooklyn featured the vast majority of the complaints and therefore would not produce a meaningful data visualization. Therefore, it was decided to aggregate the data by zip code.     

The next set of visualizations continue to explore the number of elder abuse complaints by zip code, but also introduces a third variable designed to represent which month had the most complaints reported.  Again, the first visualization relies on a bar chart to display which zip code had the most reported elder abuse complaints.  However, it also includes the complaints aggregated by month, which is indicated by the different colors now being used to create each separate “bar” within the bar chart.  Upon inspection of this visualization, it was determined that it was difficult to interpret any potential meaningful information when considering complaints made by month.  As a result, it was decided that aggregating by zip code might be too narrow of a constraint, and a second bar chart was created to explore the total number of elder abuse complaints reported each month based on the entire data set.  Broadening the parameters of the visualization, and aggregating solely by which month the complaint was reported, made the bar chart easier to interpret and possibly revealed a trend that fewer elder abuse complaints are made during the summer months.  However, this finding should be interpreted with caution because the data set extends from 2017 to the present and 2019 has not yet completed.  Despite the cautious interpretation, these two visualizations reveal how data sets influence and drive the iterative process of data visualization when hoping to present data in a meaningful way. 

In an attempt to further understand when elder abuse complaints were being reported, it was decided that two additional bar charts would be created.  The first (top) bar chart is the same from the previous set of visualizations and displays elder abuse complaints by month.  The next bar chart displays elder abuse complaints by day of the week and the third (bottom) bar chart displays elder abuse complaints by hour of the day.  Taken together, it appears as if the majority of the elder abuse complaints are made Monday to Friday between the hours of 8:00 A.M and 5:00 P.M. 

Discussion

The data visualizations created from the current data set exploring NYC 311 elder abuse complaints reveal a number of important findings.  First, despite the lack of a clear definition of “elder abuse,” the fact that elder abuse complaints are being reported within the City of New York should be of concern to researchers, health care providers, government officials, care takers, and all residents of New York City.  Abuse in any form is a traumatic experience that frequently results in severe and persistent physical and mental health consequences.  Second, it appears as if certain zip codes are more likely to report elder abuse versus other areas of the city.  Brooklyn in particular produced the vast majority of the elder abuse complaints and these initial findings raise questions about that borough and the potential well-being of the older adult residents who live there.  Based on the current data set and visualizations it is difficult to determine what might be causing certain zip codes to be over represented, but it nonetheless provides at least initial geographical evidence of elder abuse complaints.  Finally, the visualizations would also suggest that within the zip codes making elder abuse complaints, it appears that day of the week (Monday – Friday) and time of day (8:00 A.M. to 5:00 P.M.) might also be implicated in elder abuse reports.  Once again, the data set fails to address causal factors driving these trends, but it does suggest the importance of NYC 311 being available/operational during “normal” working hours and days of the week when considering elder abuse complaints. 

On the surface, it seems as if certain trends have begun to emerge based on the data visualizations presented within the current project.  However, it is worth noting that the visualizations included within this project potentially yield more information about how NYC 311 is utilized versus the prevalence and incidents of elder abuse.  That is, the data visualizations perhaps reveal more information about NYC 311 reporting habits as opposed to specific information regarding elder abuse.  Although it is important for the New York City Department for the Aging to understand where elder abuse is occurring, the visualizations appear to primarily suggest which areas of the city (zip code) are the most likely to use NYC 311 to report elder abuse.  In other words, the lack of elder abuse complaints from other boroughs (Manhattan, Staten Island, etc.) is not an indication that elder abuse is not occurring in those areas, it simply means that residents in those areas have not utilized NYC 311 to make elder abuse reports.  Likewise, the lower number of complaints reported on the weekends and prior to 8:00 A.M./after 5:00 P.M. should not be interpreted that elder abuse does not occur during those time periods, but rather residents are most likely to report elder abuse during the “normal” working hours/days of the week.  Based on this interpretation, the visualizations might suggest that Brooklyn residents are the most likely to make elder abuse reports during “normal” working hours through NYC 311 as compared to their fellow New York City residents in other boroughs.  Thus, from the perspective of resident utilization, the current data visualizations seem to support previous findings that suggest an unequal use of the NYC 311 program by the residents of New York City (Kontokosta, Hong, & Korsberg, 2017).  Therefore, any conclusions about elder abuse specifically should be interpreted with caution, and that any subsequent initiatives, policies, or decisions regarding the prevention/intervention of elder abuse should not be based solely on the current NYC 311 data set. 

Limitations and Future Directions

The current data visualization project is not without limitations.  Since the data visualizations are driven by the underlying NYC 311 data set, a limitation of the current project would be the lack of information needed to draw a stronger set of conclusions.  For example, the NYC 311 data set did not include demographic information or latitude and longitude information.  Descriptive statistics would add further texture and meaning to the data visualizations.  For instance, if hoping to make policy decisions about preventing elder abuse, it might be useful to know “who” is making the elder abuse complaint.  While it is perhaps reasonable to assume that most New York City residents are not calling in “noise” complaints on themselves, it remains unclear if elder abuse complaints are being reported caretakers, family members, or perhaps by the older adults who are experiencing the abuse.  Additionally, the absence of any latitude and longitude information could be considered an important limitation.  This information is not only useful from a visualization standpoint, but it prompts questions about the consistency of the data set based on the fact that other NYC 311 complaints, such as “noise,” have included latitude and longitude information.  Another important limitation would be that the data set and associated Data Dictionary (code book) did not provide an operational definition of “elder,” “abuse,” and/or “elder abuse.”  The lack of clarity surrounding these terms makes it difficult to draw conclusions about elder abuse from a research, health care, or public health perspective.  Additionally, it should be mentioned that the current data set was filtered by the date of when the elder abuse complaints were reported, and therefore only contains data from January 2017 to the present (September 15, 2019).  A more comprehensive data set might have produced more compelling findings. 

Finally, the data visualizations included within this project only reveal potential correlations between the variables in questions.  As a result, any inferences of causation would be inappropriate.  In addition, any perceived correlations, trends, or differences regarding the elder abuse complaints have not been tested for statistical significance.  In other words, the data visualizations within the current project may reveal relationships amongst the variables that are simply due to chance.  Therefore, any findings about elder abuse specifically, or the habits of NYC 311 users, should be interpreted with extreme caution.  In turn, future studies should seek to gather additional information in regards to location, relevant demographic information, and operational definitions of the terms under examination.  Additionally, a larger data set might yield potentially more useful findings, especially when combined with tests for statistical significance. 

References

American Psychological Association. (2019). Older Adults: Health and age-related changes. Retrieved from https://www.apa.org/pi/aging/resources/guides/older

City of New York. (2019). About NYC 311. Retrieved from https://portal.311.nyc.gov/about-nyc-311/

City of New York. (2019). NYC Open Data: 311 service requests from 2010 to present. [Data file and code book]. Retrieved from https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9

City of New York. (2017). NYC Open Data: Overview. Retrieved from https://opendata.cityofnewyork.us/overview/

New York City Department for the Aging. (2019). Commissioner’s Message. Retrieved from https://www1.nyc.gov/site/dfta/about/commissioners-message.page

New York City Department for the Aging, Age-friendly NYC Commission. (2019). Age-friendly NYC: New commitments for a city for all ages. Retrieved from https://www1.nyc.gov/assets/dfta/downloads/pdf/publications/AgeFriendlyNYC2017.pdf

Kontokosta, C., Hong, B., & Korsberg, K. (2017). Equity in 311 reporting: Understanding socio-spatial differentials in the propensity to complain. arXiv preprint arXiv:1710.02452.