Data Science meets
Full-Stack Execution.
Diego Villagran Salazar — Data Scientist & Full-Stack Developer.
I build machine learning systems, analytics products, and scalable web apps that create measurable business impact.
Featured Projects
Selected systems // 2024—2025
Transportation stakeholders needed reliable insights across Uber and Lyft trip patterns, pricing, and airport operations in New York City.
Built an interactive Streamlit analytics platform with predictive ML models, geospatial maps, and multi-tab operational dashboards.
Delivered fare prediction with R² > 0.85 and airport classification with 92% accuracy for practical decision support.
Raw environmental data from hundreds of IoT sensors was fragmented and difficult to convert into policy-ready insights.
Designed a cloud ETL architecture using Azure Databricks, PySpark, PostgreSQL, and BI reporting for continuous analytics.
Processed 2M+ daily records from 500+ sensors and transformed noisy streams into consistent, actionable health indicators.
Students needed a more engaging and structured way to practice coding with feedback and measurable progress.
Developed a gamified full-stack platform with Astro/Django architecture, secure auth, and learning-oriented UX.
Scaled to 10k+ users and earned 2nd place in the 2024 EdTech Innovation Awards.
Systems & Capabilities
Machine Learning Pipelines
From preprocessing and feature engineering to training, evaluation, and deployment of predictive models.
Data Engineering
ETL orchestration with Python, PySpark, SQL, and cloud platforms for reliable high-volume analytics workflows.
Analytics Products
Interactive dashboards and decision systems with Streamlit and Power BI focused on real-world business metrics.
Web Platform Development
Scalable full-stack applications with Next.js, React, TypeScript, and cloud-ready deployment practices.
Working Principles
“Turn complex data into clear decisions and scalable products.”
01. Impact over output
I prioritize measurable outcomes: model accuracy, decision quality, processing speed, and business value.
02. End-to-end ownership
I build complete systems, from data collection and cleaning to production deployment and monitoring.
03. Clarity at scale
Good architecture keeps complexity contained so teams can iterate quickly without breaking reliability.