From the course: Data Science Methodologies: Making Business Sense

Unlock this course with a free trial

Join today to access over 22,400 courses taught by industry experts.

Solution 2

Solution 2

(upbeat music) - [Instructor] Here is my solution to the second challenge to deploy the car evaluation MVP on Heroku server. Let us first look at the steps. We start with working on the model because that needs to be deployed first. We need to install Gunicorn, create procfile and requirements.txt. Then we need to pull in data inside the repo. Of course, this would depend upon whether you want to train the model on Heroku or elsewhere. If elsewhere, then you will need a pickle file which you will use instead of training the model in your code. Then you push your code to Git, then go to Heroku to create the pipeline with an app that will connect to the Git repo for the model, and finally build and deploy. Now you will have a URL for the model's API, which needs to be plugged into the web application. And then once again, prepare the code with Gunicorn, Procfile and requirements.txt, push it to Git, create another…

Contents