Last GHW I attended a series of lectures on Exploratory Data Analysis (EDA). This looked like a good opportunity to perform that.

What it does

Used python language to construct a series of graphic representations of 279K calls to the Tucson Police Department in the year 2022.

How we built it

Downloaded the Data from the City of Tucson website. Constructed a VScode python project to provide a series of statistics on the count and uniqueness of columnar data. Used both matplotlib and seaborn as well as pandas and numpy to eliminate rows of data that had missing data.

Challenges we ran into

Data integrity was found to be complete and concise but fairly cryptic as it utilized terminology specific to the Police Science data domain. I did not understand the Automatic calling log process and still wonder a bit about why the data showed some of the patterns that it did. This could probably be quickly resolved with a call to the local cities office of data integrity.

Accomplishments that we're proud of

Identified the months that had the fewest police/citizen interaction as well as the days.

What we learned

In 2022 the TPD was experiencing reductions in the number of employed police officers, yet was still able to respond to >279K calls. This is commendable in a city of >1 million population. Looking at the Days of the week chart shows that there is a definite DOW pattern to calls and responses. (Lower on weekends). The population of the city increased by several hundred thousand in the winter months, yet the call and response pattern falls in this period and increased as the weather became hotter with only the full time residents in the city. (This means the part time population known as 'snowbirds' is an older and much more law abiding sector of the population). The only anomaly is the month of January where the increase in crime/response is due to the new year festivities ie. drug/alcohol abuse etc.

What's next for Data Visualization for Police Calls 2022 - City of Tucson

Some exploration of the data in Geocoding format for spatial identification of crime/response patterns.

Built With

Share this project: