A Multimodal Approach for Crime Pattern Analysis

Crimes occur everywhere around us, impacting a society's economic development and quality of life. Despite the fact that crimes happen everywhere, thieves choose to target well-known locales whenever there is a chance. In order to detect a pattern among these crimes, it is, therefore possible to process data such as the crime type, date, area, and so on using a data mining approach. This might make it easier for us to determine whether a place is safe or harmful at a specific time, day, and location. Through this, we seek to increase people's awareness of their surroundings, assist them in staying alert at risky times, and encourage them to take safety precautions. Additionally, this could be helpful to the local police force so they can send a proper police force and ensure high securityin crime-prone areas and times. Our data mining project intends to conduct tactical crime research on a dataset of actual crimes that occurred in Los Angeles, California, between 2010 and 2015. We employed a variety of techniques and algorithms to build various graphs to help us assess the data and be able to forecast the crime rates in particular regions at a given time. These graphs allowed us to extract common patterns of crime from the dataset. We have included demographic data along with statistical analyses of various crime kinds.