As a data science noob, I'm currently on the hunt for the most efficient learning path. Creating a project is highlighted as the most obvious step. However, yet I think I've learned the minimum enough to get started with it,  Something has had stopped me for a while, so I found the best way possible to keep on track: document it, I guess I'll just see by myself if that's true or not. But, for now, this blog will be an up-to-date journey of my process/feelings throughout developing my own first ML project.

When you start, everything comes down to a question: What am I going to do? Is this idea too much? Is it enough? For me, the best way possible to cope with that was simply not to think about it. I found a Twitter thread from @Svpino where he talks about several ml projects (I'm going to leave it down below) and I just picked the one which came to me the most: Music player. However, I wanted podcasts, why not?  a Podcast player sounded great!

In the same threat, Alejandro (@Svpino) talks about a fascinating approach: try to divide your project into features. Later, try to find the easiest way possible to accomplish every single one

This is what I thought the project should have:

1.  an interface

2. something to decide what podcast to recommend 3.  Users should be able to actually listen to the podcast

As my first approach, I think I will use simple HTML CSS, and JS to create the interface. I found as well a dataset from Kaggle about podcast reviews and I think I will use it with a classifier to decide whether people like it or not, then I'd rank it and add it to a recommendation page. Probably divided into themes as well, this is the simplest way I can think of. Lastly, I can think that the best way to listen to the podcast is by being redirected to Spotify.

Nonetheless, I have plenty of ideas I think I would add later. For example, firstly, I would like to use my own dataset to recommend the trendiest podcasts. Then it would be amazing to use sentiment analysis to recommend in a better way, even face detection, to see how you are feeling today!

I'm not gonna lie, I'm feeling very happy about starting a project at the moment, everybody says that the emotional part of creating a project is a bit of a slippery slope. Consider myself as your guinea pig documenting its feeling journey to see whether it is right or not.

BTW, this is the Twitter threat :)