If there's one thing that gets everyone stoked on AI it's Deep Neural Networks (DNN). From Google's pop-computational-art experiment, DeepDream, to the more applied pursuits of face recognition, object classification and optical character recognition (aside: see PyOCR) Neural Nets are showing themselves to be a huge value-add for all sorts of problems that rely on machine learning.

Optical Character Recognition is an old and well studied problem. The MNIST dataset, which comes included in popular machine learning packages, is a great introduction to the field. In scikit-learn, for instance, you can find data and models that allow you to acheive great accuracy in classifying the images seen below:

Scraping data from websites is one of those things that sounds really cool and high-tech, but it's pretty simple and fun. In case you're too lazy to read the docs for Beautiful Soup and Requests, here's a quick and intro to webscraping in Python. The most basic functionality is demo'd below:

If you've ever ridden a New York City subway, you've seen that there are a lot of people that use these things. It's not hard to see why: they're dirty, hot, jam-packed with pissed off commuters.. wait.. actually it's that they're the only way to get around here that doesn't risk you getting rammed into the guardrails of the BQE.

So here's my blog.