There’s a joke on the Internet about how most Machine Learning Models end up in a Powerpoint presentation. If you are a data scientist, probably you can’t deny that fact. But things have started changing, especially with solutions like Voila, Streamlit, Dash by Plotly. These are part of a niche within ML and Data Science community enabling Data scientists and Machine learning Engineers — who’s no Web Developer — to develop and deploy Machine Learning Applications (as Web apps — like SaaS).

ML Web Apps

Their main pitch is that you can write web applications simply with your familiar syntax of Python (that means, no javascript required ). As this sort of ML Web App development is kicked off, many data scientists are still used to and like to code on Jupyter Notebooks, unlike Web Developers/Software Engineers who mostly prefer a full-fledged IDE like VSCode or Sublime Text. That’s also the case with Hobbyists who use Google Colab to develop such Machine Learning Web App Prototypes — either as a hobby project or as an assignment for a potential job opportunity. Ironically most of the time, those things are shared as Jupyter Notebook or screenshots or someone has to take the pain of deploying them in pythonanywhere or Heroku

ngrok

If you are in such a situation, This hack would come very handy because you’d be able to share your ML Web Application using a Public URL with anyone in the world(temporarily). This hack uses a very popular secure tunneling solution called ngrok You can read more about what ngrok is and what it does here.

#web-development #data-science #google-colab #data analysis