Hi Karthik — thanks for the feedback. I’ve made some updates to the code since I wrote this article to clean it up and make it more flexible. I’ve also applied it to an open-source dataset so that somebody can recreate the process from end-to-end. This repo https://github.com/ryanmark1867/ai_conference_june_2019 includes the code, instructions on how to use it, plus intermediate datasets as pickled dataframes to make it easy to run subsets of the whole flow.

To get back to your original question, the dataset used in the repo above includes a several numerical features. Look for the definition of the get_model() function in the streetcar_DL_train-trimmed.ipynb notebook and you will see how the layers are built up for categorical, text and continuous (AKA numerical) columns.