communities-and-crime

Author: Michael Redmond (redmond 'at' lasalle.edu); Computer Science; La Salle University; Philadelphia, PA, 19141, USA Source: [UCI](https://archive.ics.uci.edu/ml/datasets/Communities+and+Crime+Unnormalized) - 2011 Please cite: U. S. Department of Commerce, Bureau of the Census, Census Of Population And Housing 1990 United States: Summary Tape File 1a & 3a (Computer Files), U.S. Department Of Commerce, Bureau Of The Census Producer, Washington, DC and Inter-university Consortium for Political and Social Research Ann Arbor, Michigan. (1992) U.S. Department of Justice, Bureau of Justice Statistics, Law Enforcement Management And Administrative Statistics (Computer File) U.S. Department Of Commerce, Bureau Of The Census Producer, Washington, DC and Inter-university Consortium for Political and Social Research Ann Arbor, Michigan. (1992) U.S. Department of Justice, Federal Bureau of Investigation, Crime in the United States (Computer File) (1995) Description The source datasets needed to be combined via programming. Many variables are included so that algorithms that select or learn weights for attributes could be tested. However, clearly unrelated attributes were not included; attributes were picked if there was any plausible connection to crime (N=125), plus the crime variables which are potential dependent variables. The variables included in the dataset involve the community, such as the percent of the population considered urban, and the median family income, and involving law enforcement, such as per capita number of police officers, and percent of officers assigned to drug units. The crime attributes (N=18) that could be predicted are the 8 crimes considered 'Index Crimes' by the FBI)(Murders, Rape, Robbery, .... ), per capita (actually per 100,000 population) versions of each, and Per Capita Violent Crimes and Per Capita Nonviolent Crimes). A limitation was that the LEMAS survey was of the police departments with at least 100 officers, plus a random sample of smaller departments. For our purposes, communities not found in both census and crime datasets were omitted. Many communities are missing LEMAS data. The per capita crimes variables were calculated using population values included in the 1995 FBI data (which differ from the 1990 Census values). The per capita violent crimes variable was calculated using population and the sum of crime variables considered violent crimes in the United States: murder, rape, robbery, and assault. There was apparently some controversy in some states concerning the counting of rapes. These resulted in missing values for rape, which resulted in missing values for per capita violent crime. Many of these omitted communities were from the midwestern USA (Minnesota, Illinois, and Michigan have many of these). The per capita nonviolent crime variable was calculated using the sum of crime variables considered non-violent crimes in the United States: burglaries, larcenies, auto thefts and arsons. (There are many other types of crimes, these only include FBI 'Index Crimes') Some further pre-processing of the dataset must be done. Choose the desirable dependent variable from among the 18 possible. It would not be interesting or appropriate to predict total crime (e.g. violent crime) while including subtotals (e.g. murders) as independent variables. There are also identifying variables (community name, county code, community code) that are not predictive, and would get in the way of some algorithms. Weka's Unsupervised Attribute Remove Filter can be used to remove unwanted attributes. The FBI notes that use of this data to evaluate communities is over-simplistic, as many relevant factors are not included. For one example, communities with large numbers of visitors will have higher per capita crime (measured by residents) than communities with fewer visitors, other things being equal.