It's hard to say what brought about Octopus. Why Octopus? A sea creature hardly seems appropriate for a project such as this. But beyond the surface, a creature as slimy and mysterious as the legendary "donkey of the sea" is perfect for a project like this. In a way, we aspire to be like the octopus, whose brain to body mass ratio is the highest of all invertebrates. Also, tasty.
What it does
Octopus first prompts the user for their location. Using this and date/time data, Octopus uses a model constructed from hundreds of thousands of crime reports from around Chicago to determine which type of crime is most likely to occur at that location and time. Although there are heat maps of cities detailing dangerous and safe areas, none of them explicitly state the chances of certain crimes occuring.
How I built it
The core of this program runs a neural network in Python, as well as a heuristic we wrote to determine how dangerous an area is and what crime is most likely to occur in the area. The website that hosts it uses HTML. Part of the challenge was formatting the crime data (csv) into a usable form, which thankfully is not too difficult in Python.
Challenges I ran into
We're monkeys who make tons of syntax mistakes, spellig mistakes, logical mistakes, basically every mistake a CS student could possibly make.
Accomplishments that I'm proud of
It's machine learning! Though it's just a start, we're all glad that it runs properly. Considering it's also a novel idea, we're proud of what we've done here, especially since we did it in 24 hours.
What I learned
Crime rates spike around New Year's, especially sex offense :( On a cheerier note, I've learned that ACM is awesome and that I need sleep to properly function as a human.
What's next for Octopus Crime Check