When I want to learn a new topic, I immerse myself in the writings of those who have generously helped me understand various subjects. Writing a blog is my way of paying it forward and giving back to the community that has supported me in my own learning journey. By sharing my knowledge through articles, I hope to contribute to the collective pool of information and provide others with the same kind of guidance and enlightenment that I have received.

Nowadays studying Data Science and develop personal projects is easier than ever. There are many websites with free and interesting datasets out there. This is just one of many examples of that. At the Barcelona government website, I found this dataset that gives us a very good starting point for a Time Series analysis. Continue reading…

Last summer I had what I thought was a fantastic idea: let’s code a two players version of the game Asteroids, using Pyscript, and then use reinforcement learning to teach a neural network to play that game. And yes… “I thought it was a fantastic idea”. Well, it wasn’t that fantastic! First of all, Pyscript is a relatively new library, with hardly any documentation. Secondly (and most important), Pyscript is based on Pyodide, which doesn’t support any Deep Learning library such as Keras, Tensorflow or Pytorch. And I discovered this… after two months of coding the game itself! Continue reading…

This article will go through a basic implementation of a neural network using only NumPy. It will be divided in two, one part for the forward pass and the other (here) for the backward pass. If you want, you can download the code in this repository, and read the Jupyter notebook version of this same article. Continue reading…

Beatiful Soup (in its latest version, BS4) is a well known and wide used Python library, very useful to scrape data from websites. Although there are more advanced tools, BS4 is still one of the first choices among data analysts when they need to gather information. Continue reading…