Correlation of drug/narcotics and prostitution related crimes in San Francisco.
Crime is a complex and multifaceted phenomenon that affects individuals and societies in numerous ways. From petty theft to violent crimes, criminal activities are a major concern for law enforcement agencies, policymakers, and communities around the world. While each type of crime is unique in its nature and impact, there are often correlations between different types of criminal activities.
Understanding the relationships between different types of crimes is essential for developing effective crime prevention strategies and policies. For instance, studies have shown that individuals who engage in petty theft or property crimes may be more likely to progress to more serious crimes, such as burglary or robbery. Similarly, drug-related crimes and prostitution often go hand-in-hand, highlighting the need for a comprehensive approach to address these issues. As technology advances, researchers, policymakers, and law enforcement agencies have new tools to study patterns in criminal activities and effectively combat crime.
In this article, we wanted to explore the correlations between drug and narcotic consumption and prostitution, following the conclusion of various papers and books on the topic. The book “Prostitution and drugs” and the papers “Prostitution, Drug Use, and coping with psychological distress” and “Pathways to prostitution: The chronology of sexual and drug abuse milestones”, inspired us for our exploratory data analysis.
We wanted to highlight the similarities and differences between these two forms of criminal activities. The data provided for our study is the “Police Department Incident Reports: Historical 2003 to May 2018” of the city of San Francisco (USA). We will examine various aspects of the evidence reflected on the data as a first approach to identify factors that contribute to correlation between these types of crimes.
How have the number of crimes changed over time?
Firstly, we analysed the time series of both crimes from year 2003 to 2017. In the following plot you can choose and look into the annual, weekly or hourly trends. Note that the values in the number of crimes have been normalised in order to have a better comparison of the two crime types.
Insights and conclusions
Hours per day:
Initially, one may not be able to see any correlation between the two types of crimes. However, at a closer look and understanding of the graphics, we can see that drug/narcotic consumption related crimes start increasing from 7am until 6pm, and we can see the prostitution starts increasing at 11am and has its peak at midnight. This means that there is the same trend in both type crimes with a 4-6 hour difference.
Days per week:
In the weekly plot we can see that in both types of crime, there is an increase of occurrences during the weekend. But there is a high number of prostitution crime incidents on Mondays. Is this really something unexpected? Taking into consideration the fact that the hourly plot reflect that the peak of incidents occur at midnight, we can infer that the trend of Sunday evening continue up to Monday morning.
Over the years:
The data that we are managing contains crime incidents from 2003 until 2017. We can easily observe that in both crime types, the general trend in number of incidents decreases over the years.
Where did the crimes occur?
Another important aspect to include in our analysis is the spacial crime occurrences. We show in the following plot a map of San Francisco in which we show the most frequent spots where both drug consumption and prostitution where committed, the number of reported cases is shown after clicking a marked point.
The map above displays eight locations with the highest reported occurrences of prostitution and drug crimes. It is noticeable that these crimes have their highiest frequency in two main areas. Even though both types of crimes happened in both areas, the exact places were not the same.
What is interesting is that in both areas, the places with the most crime for each type were really close to each other. This observation raises the question of whether these occurrences are the result of intentional planning rather than coincidence.
Using maps to study crime patterns is important because it can help police focus on areas with higher probability of crime occurence. This might help them prevent crime from happening in the future.
Can we predict the future occurence for the crimes?
To better understand the trends in the number of crimes committed over the years, we used linear regression. Note that in the plot with both crimes, the y-axis scale is the same, so that the visualisation of the linear model is not misleading and can be compared easily.
It is evident from the two regression plots that there has been a downward trend in the number of these crimes. Both regression slopes demonstrate a negative inclination, indicating a decrease in the occurrence of these offenses over time.
The slope of the model describing Prostitution is -99.375. This corresponds to a decrease of 78% in this type of crime during the 15 years of our study. For the drug/narcotics related crimes, the model shows a slope of -504.393, which means that from 2003 until 2017 (15 years) there was a decrease of 63% in drug/narcotics related incidents.
However, it is important to note that a simple linear regression model may not be the most comprehensive or accurate method for capturing the complexities of crime trends. Consequently, we cannot make reliable predictions solely based on these findings. To draw more robust conclusions and inform potential policy decisions, we recommend conducting further analysis, employing advanced statistical methods, and considering additional variables that may influence the prevalence of drug-related and prostitution crimes in San Francisco.
Conclusion
Understanding the correlation between different types of crimes is critical for developing effective crime prevention and intervention strategies. By this visual study of the data of previous years, we have tried to show a first approach to this complex problem. For a further analysis, it would be beneficial to take more variables into account for analysis, for example personal information about the person who committed the crime, such as gender, age or race.