Skip to content


The Dashboard

Users may manipulate data visualizations through a number of filters that appear to the right of the dashboard (e.g. date, victim age, victim sex/gender, etc.)

Layers of geospatial information about population density, racial makeup, and median income were created by students in Professor Megan Heckert’s course on “Applied Analysis with GIS” at Swarthmore College using demographic data from the 2012 American Community Survey 5-­year estimates by the US Census Bureau.


The Delaware County Homicide Database displays data trends for all homicides since 2005, even though we have been especially interested in gun-related homicides. The data does not include other forms of gun violence, such as injuries, or suicides. Victims constitute the unit of analysis.

To create the database, we relied primarily on two kinds of sources: law enforcement data and journalistic sources.

We have relied heavily on the Pennsylvania Uniform Crime Reporting System (PA UCR), part of the FBI’s national Uniform Crime Report system. That data is now reported through the National Incident-Based Reporting System (NIBRS). (Note: 2019 data draws on the PA UCR and Summary Reporting System data from the Return E Homicide Report.) The PA UCR provided data tables for a number of variables: homicide type, homicide situation, victims (#), victim race, victim ethnicity, victim age, victim sex, offenders (#), defendant race, defendant age, defendant sex, relationship type, death circumstances, whether a homicide was deemed justified, weapon type, and whether the weapon was automatic. As our news sources did not always report incidents with the same consistency in their terminology and information as PA UCR, the latter information was particularly crucial in providing a baseline of information for incidents.

The Delaware County Daily Times served as the journalistic foundation for our database by providing a list of each homicide in Delaware County from 2005-2017. For each incident, the list included the name of the victim, date of the incident, and city/township/borough in which the incident occurred. From these lists, we searched for news articles written about incidents from sources including but not limited to The Delaware County Times, ABC news station, CBS news station, and the Philadelphia Inquirer.  If a homicide received journalistic coverage, the approximate location was usually reported (most often a city block or nearest street intersection).

The PA UCR data tables did not record victim names, so in order to connect instances recorded in the UCR to corresponding homicides with geographic locations from our study of journalistic sources, we matched the date of each incident, victim’s race, and victim’s age. If a homicide received journalistic coverage, the approximate location was usually reported (most often a city block or nearest street intersection).

While most homicides were reported to some extent in the Delaware County Daily Times or other outlets, many offered only brief stories with little detail. Furthermore, these stories were often published soon after the event, which often did not leave time for further clarification of the initial report, as additional information might have become available through police or medical examiner investigations. We occasionally encountered disagreement between news stories and the PA UCR. Some news stories about a homicide had no match in UCR data, and, likewise, some UCR homicides did not match any news stories. Aggregated data available to users via the dashboard on this site in most cases represents information from the Uniform Crime Report. When a homicide incident was reported in the press but did not appear in the UCR, we gathered information we could glean from journalistic reporting. Thus, the information presented here does not exactly represent the results of the Uniform Crime Report. Numbers of homicides usually match annually with a difference of no more than five cases.

Data in the Delaware County Homicide Database will be cleaned further and improved over time. Thus, small fluctuations may appear in dashboard reports.