Heroku-End-To-End Machine Learning Project With Deployment

In this data science machine learning project tutorial, we are going to build an end to end machine learning project and then deploy it via Heroku.

Feel free to test the App here

JOIN TELEGRAM GROUP

I am writing this article because with my current role as a Data Science Product Manager and Consultant for multiple companies I interview almost 25 Data Scientists every month for my company and other companies I help to grow. During the interview process, I see candidates who are really interested to join the company but they don’t have anything to prove their desire or confirm their skills. Sometimes it pains me to reject them but I cannot recruit someone based on just a mere desire I sense. Having the desire is the number one thing I look for but I need proof to justify that desire. I, therefore, give preference to those who have some portfolio to show.

Data Scientist is an incredible career to have. At least a good career to survive the next career shift. You will agree with me that getting a Data Science job is not a piece of cake considering a base salary of $120,000, that’s good enough to make your life heaven on earth.

But how do you get in as a FRESHER?

I will show you how soon. be patient with me and keep reading…

Let’s see some of the Job Descriptions of a Data Scientist

Data Scientist job description

These are just two of the screenshots I took on Glassdoor for a Data Scientist job description. You can just do a google search and see the requirement as well. These skills are not impossible to achieve but they are not easy to have them as one person unless you add some little bit of witchcraft..lol. You can build it gradually but you have to be smart and build it the right way else it will take you years. The most important of all is the foundation. If you get your foundation right, the rest are just add-ons.

My advice: start with a course that focuses on teaching and building end-to-end data science projects. By end-to-end, I mean from data cleaning to deployment.

Before we start, please note that this tutorial is part of the Data Science Project Mastery Program, which is a practical hands-on data science tutorial for anyone to learn data science right from the basics to advance by building projects and great Data Science Portfolio. Feel free to check it out.

What you can do is to build a Data Science Portfolio. Pick up a problem statement and build it end-to-end. From data cleaning to deployment. Don’t just do it, write the process down, which helps the recruiters to know what is involved in the project. It also serves as a reference to some of the concepts that you might forget later on. Publish it to increase your visibility. I have hired 4 Data Scientist on Medium simply because I saw their article about a project they have done and I was convinced that they can be resourceful if they get the chance, although I know they might have copied the code or the solution from somewhere else, but I also know they can find their way out with a little push.

If I want to employ a Data Scientist, I want to see something. Not necessarily a superhero but someone who has something to show that yes I can find my way out if you give me a task to do because I have done something related to that before.

Let me just pause here and ask you. Why do you think employers ask for work experience?

The answer is simple: They want proof of concepts.

So if you are looking forward to starting your journey as a Data Scientist, make sure you have proof of concept by making your Data Science Portfolio ready.

Am just going to show you a simple project that you can follow along to do other projects and even highly intuitive ones.

Over here, we will be predicting the prices of the various airlines in India.

Let’s do this step-by-step

1. Choose your use-case
2. Get your dataset prepared and build your model and choose the best model
3. Create a web App using Flask
4. Commit your code to Github
5. Create your account on Heroku
6. Link your Heroku account to your Github account
7. Deploy and test your App

Let’s get started !!

In this Data Science Project, we will mainly be using python for data science. If you are not familiar with python or you want to brush up your skills on Python, you can use this video where I have explained every concept you need to know about Python for Data Science.

If you are here, I believe you are familiar with Python and we can proceed.

>>

The first step we are going to take is to choose our use case. In this scenario, we will be predicting the price of the various airlines in India as mentioned earlier. Feel free to choose your own use case.

First of all, let’s get the files ready. You can download these files and work along with me.

Tools Needed:

1. Jupyter Notebook OR Google Colabs
2. Pycharm(make sure to download the community version for free) OR any other IDE of your preference
3. Flask
4. Heroku

Files Needed:

1. Prepared Notebook

All File Downloads

IMPORTANT !!!!!

For Flask and API concepts, I have explained in a previous video about Zomato(an Indian restaurant aggregator and food delivery start-up founded by Deepinder Goyal and Pankaj Chaddah in 2008), which you might want to watch that before moving on.

Watch the ENTIRE PLAYLIST HERE

==========

I hope you have watched the recommended videos on the previous project and enjoyed it.

Alright, now that we are on the same page, let’s proceed…

In order for you to get a better understanding and make the best out of this tutorial, I have made a video of everything you need to know so that you can better follow along.

Watch the ENTIRE PLAYLIST HERE

If you are interested in learning Data Science, the right way, I advise you to start with this course that focuses on teaching and building end-to-end data science projects.

I will also be posting more FREE Data Science Projects on the Youtube Channel, so don’t forget to subscribe and hit the notification bell to get notified anytime I poste them.

And hey! don’t forget to like this article.