I just wanted to document my first interactions with ChatGPT’s new Code Interpreter.

I used a data set from the Toronto Police Service Public Safety Data Portal.

This is a screenshot of what the data looks like in the CSV file I downloaded.

1. Describing the Dataset itself

I asked the Code Interpreter “Can you provide some insights about crime in Toronto based on this data?”

There is a lot going on under the hood which you take a look at, as shown below. ChatGPT uses Python code to create the visualizations.

2. Analyzing the Data

ChatGPT provides some analysis of the data.

3. Visualizing the Data

The real cool stuff when ChatGPT visualizes the data. The following were created as image files.

4. Asking ChatGPT for some Unusual Visualizations

I asked ChatGPT for some unusual or interesting data visualizations and it gave me the following;

WordCloud for Neighbourhoods: This word cloud represents the frequency of homicides in various neighbourhoods. The size of each neighbourhood name in the word cloud corresponds to the frequency of homicides in that neighbourhood. Larger names correspond to neighbourhoods with more incidents. This visualization provides an immediate visual understanding of the areas with the most frequent incidents.

Donut Plot for Crime Type: This donut plot provides a visual breakdown of the types of homicides in the dataset. The size of each section corresponds to the frequency of each type of homicide, with labels and percentages for clarity. This visualization provides a clear and immediate understanding of the most common types of homicides.

My Other Blog Posts

If you liked this you might enjoy my other blog posts.