Oloren Datathon

Twitter for the Oloren Datathon. Related to @Oloren and @OlorenChemEng. Check it out!

[Oloren Datathon @OlorenDatathon]

A few comments:

1. Super easy to improve this. Give it a shot!
2. @OlorenChemEng makes reproducibility much smoother.
3. Submit to the datathon!

[Oloren Datathon @OlorenDatathon]

Some sample code for the datathon!

Here's an example base boosted model built with @OlorenChemEng in less than 10 lines of code. Feel free to copy it and submit: https://t.co/EhJto7rB0H

[Oloren Datathon @OlorenDatathon]

Performance! tldr; we do decently well.

[OlorenChemEngine @OlorenChemEng]

Waterfall plots!

They're super useful for understanding base boosted models. It's easy to visually identify each model's iterative contribution.

Here's a simple base boosted model we made for @OlorenDatathon.

vis = oce.BaseErrorWaterfall(model)

[Oloren Datathon @OlorenDatathon]

Some important links:

Github: https://t.co/hanePkdmHa
Submission form: https://t.co/R4e9wPiorO
Leaderboard: https://t.co/TR5DUto5nv
Contact: [email protected]

[Oloren Datathon @OlorenDatathon]

We're holding an Oloren Datathon!

The goal is to use @OlorenChemEng to make the best model. The first dataset is hepatocyte clearance.

First place gets $1000! Special gift for beating our benchmark!

Full Rules: https://t.co/gZ9sazUSUC
Community: https://t.co/GbjuKNw3ve

[Oloren Datathon @OlorenDatathon]

Why are we doing this?

Our goal for these datathons is twofold: getting more people to use OCE, and achieving state of the art performance. We’re really interested in seeing what works, what doesn’t, and how we can improve. Would love everyone's input!