Crime prediction and analysis have been the backbone within which public safety is expected and
effective policies can be made. Using time series forecasting techniques such as SARIMA and Prophet,
this project describes crime forecasting in Andhra Pradesh, Telangana, and Maharashtra. The overall
analysis will be directed towards producing actionable insights for police and policymakers by analysing
historical crime data. After thorough data preparation involving cleaning and feature enrichment,
Exploratory Data Analysis (EDA) will be conducted to identify important crime patterns and seasonal
behaviours. The predictive model of SARIMA captures seasonal variations, while Prophet takes care of
the irregularities and missing data with flexibility and accuracy. A comparative assessment of these two
models would help to highlight their strength in forecasting crime trends. Crime patterns significantly
differ with each region with specific traits attributed to socio-economic and geographical aspects. Some
problems include data imbalance, scant historical data, and inconsistencies in records. But the project
highlights the application of machine learning to handle real-life problems in crime prevention. It opens
up a future direction integrating new, real-time data, socio-economic variables, and geospatial analyses
to make predictions more precise and provide actionable insights dynamically. Overall, the project
shows the promise of data-driven approaches in improving public safety and reducing crime rates.
