I've been thinking a lot about how advancements in NLP can be standardized; when building a sentiment analysis, you know for sure that other attributes than the actual sentiment can be highly interesting, such as the urgency.
Together with my team, I started a new open-source project, aiming to do exactly that. It's called bricks, and it is a composition of more than 50 open-source and modular code snippets, such as computing sentence complexities, emotionality detection and many more.
For context, the idea came up after watching the incredible talk "Inventing on Principle" by Bret Victor. The goal is to make the gap between Idea to Implementation shorter.
