Crime Mapping and Predictive Analysis of Crimes in Maryland, USA

The issue of crime has long existed, both nationally and internationally. Several research have looked at previous crimes and provided appropriate preventive measures. However, even in recent years, there has been a high prevalence of crime. Therefore, the primary objective of this study is to analyze the crimes that took place in Maryland, USA, between 2016 and 2020. After pre-processing and cleaning the dataset, numerous analysis and visualization findings were drawn. Improving public safety and crime prevention in the state of Maryland requires first developing a predictive analytic and crime mapping technology. By examining past trends and data, it was possible to ascertain the days of the week, time of the day, and months of the year that crimes were most likely to occur. In addition, the location of these occurrences and the most frequent form of crime each year could be inferred. The three primary machine learning algorithms that were employed to predict the kind of crime that would occur given other factors were K Neighbors Classifier, Random Forest Classifier, and Logistic Regression. The accuracy was moderate to low. The Random Forest Classifier, Logistic Regression, and K Neighbors Classifier displayed an accuracy of 52.55%, 43.86% and 38.39%. Nevertheless, by displaying past crime data trends, law enforcement officials can identify crime hotspots, identify emerging trends, and allocate resources efficiently. This would enable law enforcements to anticipate and prevent crimes before they occur, allowing for proactive interventions.