Portfolio

Apart from my main research/teaching activities (see here), I have done some additional work. At of this writing (April 2017) most of it comes from small hobby projects I have done in the area of Machine Learning and Data Analysis.

Keep in mind that this is a work in progress so expect more in the coming months!

Projects in progress

Some of these projects contain material that is intended to be published in journals/conferences and thus I cannot show anything.

Click on the image on the left to see the GitHub repository if available.

House Prices Prediction
Using different features and XGBoost, we want to predict the housing prices of a given area. Dataset taken from Kaggle which comes from the Ames Housing Dataset.
Escape of ionizing radiation in Galaxies
Using Machine Learning techniques, we want to predict the escape fraction of ionizing radiation from far away Galaxies. It is intended to be used with the James Webb telescope in 2018. Some previous work (not mine) can be read here.
Coins recognition
Using a photo from a mobile phone, we want to count the amount of money in it by using Convolutional Neural Networks and GPU Computing.

Finished projects

Click on the image on the left to see the GitHub repository

Stock Market 1-step Predictor
A Stock Market one-day predictor that takes as data Google’s stock taken from Quandl. It compares both a linear filter approach with a Kalman Filter with a recurrent Neural Network. As a bonus, it has a small investing simulation.
Anime Recommender
Japanese anime recommender written in Python. It uses K-Means clustering and previously watched series as a base for the recommendations. A nerdy application of Machine Learning. Dataset taken from Kaggle
Titanic Survival Analysis
An analysis about Titanic survivor data using Artificial Neural Networks using a few features as predictors of survival. Reasonable performance was achieved. Dataset taken from Kaggle
H1B Visa Analysis
Analysis of demographics and salaries using K-Means clustering and statistical tools. Some main points include the relationship between voters in the 2016 and the H1B concentration per capita, where most of the applications are clustered and the median salaries according to the dataset. Dataset taken from Kaggle