Hi! I'm Denilson. I recently earned my degree in Data Science from UC Berkeley, with an emphasis in applied mathematics and modeling. I'm passionate about developing predictive models, building scalable pipelines, and extracting insights from large datasets. I enjoy the challenge of transforming complex information into systems that support data-driven decisions, and I like creating a structured world that's easier for others to navigate
This is a collection of projects I completed for various purposes in different programming languages. Feel free to reach out with questions on any project :)
I self-taught HTML/CSS/JavaScript and Bootstrap Studio software in order to create my personal website.
Implementation of the BN-Grams algorithm, following a published research paper, to detect trending topics in English from multilingual tweet streams.
I self-taught some NLP theory in order to complete this project.
Profit optimization model for aircraft and crew assignments with geographical and temporal constraints
Scalable pipeline to clean, process, perform granular transformations, and interpolate sensor data using PotsgreSQL
Pipeline and machine learning model, tuning hyperparameters and enhancing quality of a noisy dataset for prediction
Extraction of predictive features using EDA to build a predictive model for house pricing
Conducted complex aggregations and data cleaning on a NoSQL data architecture using MongoDB to manage semi-structured JSON from the Yelp Academic Dataset
Utilized linear regression and sci-kit learn to classify emails based on keywords. Employed cross validation to ensure high accuracy (87.8%)
Developed SQL queries to clean and analyze the IMDb database, leveraging advanced PostgreSQL functions, regex, and relational joins to extract insights from millions of records
Optimized relational queries using PostgreSQL evaluating the cost-benefit trade-offs between various join algorithms, scan types, and indexing strategies on historical baseball dataset
email: denilsonhdez@berkeley.edu