No one are born to be data scientist, "fake it till you make it". Build neural nets with for cool problems, scikit-learn for not so cool problem and pandas for my actual work.
I move to a new blog where I can do syntax highlight properly. I wrote about stuffs occasionally.
A blog introduce a set of python tools for interactive data exploratory with examples. The visaulization below shows Country to Country donation from 1947-2013 with AidData Core Research data.Check this out
A study of Uber's approaches on how they use bayesian neural netowrk to forecast with time series data and uncertainty estimation.Check this out
A Container type classifer demo for deploying a deep learning model to a web app. You can upload a image throught the UI and it will return the prediction of the container types.Check this out
Blog post on https://codeburst.io/, 30k+ view and 2k+ likes. It summarizes and demo important tricks like magic function for Jupyter Notebook which is crucial for efficient workflow. I promote Jupyter notebook extension, which greatly extend Jupyter notebook. I also show how simple optimization can be with library like Numba and Cython.Check it out