(*Note: Data goes up to 06/25/2020)
Based on the distributions above we can infer that theft is the largest offense type committed in the Washington, D.C. area. Robbery, motor vehicle theft, and burglary are also quite high. While the more serious crimes such as homicide, sex abuse, and arson are very low for the past 10 years. Our dataset also included an additional breakout how these offenses were committed. To understand the distribution of methods used for the above incidents, let's look at an area chart...
(*Note: Left click to dive to methods, and right click to go back)
By interacting with the above tree map, we can infer that a gun or knife are not often used in incidents involving theft, robbery or burglary. Only robbery had the highest count with use of a deadly weapon (gun). However, the number of incidents recorded in this category are very small compared to the umbrella of stealing/theft type incidents for the past 10 years. To understand the distributions between each offense type, let's break them out by the last 5 years to begin to see how crime has evolved recently...
The 5-year breakout confirms our inference that theft, burglary, robbery, and just overall stealing has been the largest offense problem in the Washington, D.C. area. However, total incidents appear to be decreasing since 2016 which may point to how D.C. police have been prioritizing law enforcement resources from data-driven insights. For a better understanding how crime has changed, let's look at the entire dataset's trend of offenses for the past 10 years...
Since 2020 data is only half full of data, we omitted this year, and will only look at from 2010 to 2019. The buttons below allow users to remove an offense category for the graph to scale automatically providing a closer look of the trend. Like the above distribution of the last 5 years, the overall trend for most categories has been decreasing since 2016. However, there was a slight increase of homicide by 5 from 2018 to 2019. Since 2020 has been an abnormal year due to the pandemic, let's begin to focus on the data that is available for the year and compare using seasonality. There may be some additional trends that we may not have noticed from our treemap distribution earlier.
The above two visualizations may provide the best insight yet to our dataset so far! By interacting with the left line chart which shows how methods of offenses have changed over time, we can notice a significant decrease in the use of guns in 2016 of 2,124 incidents to 1,586 incidents in 2017. There was also a slight decrease in the use of knives. But the trend remains to be consistent so far. Therefore, we can infer that starting in 2016, the MPDC started making significant impacts in public safety and continue to do so. Both in decreasing the number of incidents relating to stealing and the use of deadly weapons.
The line chart on the right begins to show how the COVID-19 pandemic has impacted crime in the D.C. area. Since the summer and autumn seasons have not finished/occurred yet, we can remove those seasons to scale our line graph to look more closely at winter and spring. In 2019 the number of crime incidents that were recorded during the winter was 8,004 and in the spring was 7,723. Since the pandemic, the numbers have dropped to 5,490 and 5,656 respectively. That is a 37% decrease of crime in the winter and a 31% decrease in the spring. These numbers are the lowest number of criminal incidents recorded in the past 10 years! Based on the current trends and the pandemic still having a large impact to the country, it is likely that we will continue to see decreases for the summer and autumn compared to the last 10 years.
Cluster analysis or clustering is the Classification of a set of observations into subsets (called clusters) so that observations in the same cluster are similar in some way. Clustering is often considered to be a method of unsupervised learning, and a common technique for statistical data analysis used in many fields. ArcGIS is one of the best visualization tools for geographical analysis.
Since geography is a huge component to our crime dataset, let's begin to understand how crime has occurred between the different districts of the city, and what can infer from how the data has changed over time.
View Cluster Analysis