Spotify Hit Predictor
Built and ran a Random Forest model on R to predict whether a song would be a hit or a flop using a dataset of 41,000 songs from 1960 to 2019. The model got an accuracy over the test set of 85.1%
Open ProjectEduardo Mansi Medina
Recent Work
Tableau Dashboards
*Washington State Covid-19 Vaccination Analysis.
*City of Seattle Recycling and Waste Analysis.
*Airbnb Market Opportunity Analysis.
*IMDb Movie Analysis.
Craigslist scraper
Built a spider to scrap data from craigslist on cars and trucks for sale in eastside seattle. Includes multi-layer scraping to get the description of each vehicle.
Open Project
Machine Learning
Ensemble Learing:
Training several machine learning models over a synthetic moon dataset. Models include: Decision tree, Random Forest, Adaboost, Gradient booster. The models were fine-tunned by hyper-parameters using gridsearch with crossfolds.
Clustering:
Customer segmentation into different groups based on their shopping trends, Paper classification, Image compressor with K-means.
Deep Learning
The projects on Deep Learning include: Classifying reuters topics with a Full Neural Network. Image recognition for rock, paper, and scissors hand gestures with a Convolutional Neural Network (CNN). Classifying reuters topics with a Recurrent Neural Network (RNN) and Long Short-Term Memory Neural Network (LSTM).
Open Project