24h of crime in LA

3D ModelNoAI

Pro-tip: do not hesitate to pause and scroll through the animation !

Still a WIP, but I’m making progress with OpenData (after some tests with NY), and decided to explore the dataset of the arrests in Los Angeles from 2010 to present.

I animated the data by regrouping the arrests according to the time of the day (each frame represents 15 min, and the animation starts at midnight), and regrouped them under 7 distinct categories (c.f. annotation).

Some interesting patterns might be related to:

• partying and fighting in Hollywood and Downtown LA (big bright spots)
• alcohol and drug consumption late at night (red roads)
• prostitution and gambling on the boulevards (bright blue lines)
• pick-pockets at touristic places? (turquoise spots)
• …

What I’d like to try next:

• Animated legends for readability
• “Time of day” system
• Better layering: is 3D really relevant for such datasets ?