This was an assigned track, I just really love art and data science.
What it does
As of now, not a whole lot. The models that i've been able to build have been fairly inaccurate.
How I built it
scikit-learn was my first effort. I used a naive bayes multinomial classifier, but the accuracy of the model was so bad it was useless.
After that, I switched to a Gaussian classifier again running off of a naive bayes as well. The accuracy was extraordinary low. Then I switched to a random forest classifier. Again, the accuracy was low. As a result, I realized that the memory on my local machine was too small to read the images in enough detail for any good model.
Challenges I ran into
As mentioned before, memory on my local device was a huge issue. Eventually, I'll learn how to move my scripts to aws or a similar service.
Accomplishments that I'm proud of
Before this weekend I had never truly used python in its entirety for a project. This was the first time.
What I learned
Computer vision is super difficult.