This portfolio is a compilation of projects, notebooks and MOOCs completed
Spotify Music Genre Classification
This is my own end to end project using the Spotify API to predict the genre of a song/track This project consists of 2 parts: The Data Engineering and the Machine Learning/Deployment part
Part1: The Data Engineering part related to the architecture, challenges, solutions and next steps are here.
Part2: The Machine Learning/Deplyment part related to the model used, deployment architecture and next steps are here
Live version is here
Demo:
Credits:
DataTalksClub, TheSoundofAI, KodeKloud
Risk of student dropping out
Worked on a project to predict whether a student will be at risk of dropping out or not before joining JJC.
Case study of the project can be found here
Cost of brownfield in US
Worked in a team on a project to predict the cost of a specific brownfield in U.S. based on environmental, climate and demographics dataset. Project involved cleaning the datasets, presenting findings to the stakeholders, building a pipeline (using tidymodels in R) and having it ready for production following best practices of code documentation.
Case study of the project can be found here
Link: DataKind
PASSNYC Clustering Analysis (3rd place Winner for top Notebook)
One of the first Kaggle competitions in which there was opportunity to help out a not for profit company. Managed to conduct extensive EDA and gather insights using clustering methodologies. Here is the notebook
Home Credit Default Risk
Used the opportunity to go through the Applied Predictive Modelling book and apply it to the competition. The notebook got a gold rating and can be found here