Candidate: Juan Sebastian Urrea Program: PM Accelerator — AI Engineer Intern
"By making industry-leading tools and education available to individuals from all backgrounds, we level the playing field for future PM leaders." — PM Accelerator
This repository contains two independent technical assessments:
pm-accelerator/
├── weather-app/ # Full Stack Assessment (Assessment 1 + 2)
└── weather-forecast/ # Data Science Assessment (Advanced)
Location: weather-app/
A production-ready full-stack weather application built with Next.js 14, Turso (libsql), and real-world APIs. Covers both the frontend (Assessment 1) and backend (Assessment 2) requirements.
| Area | Details |
|---|---|
| Frontend | Responsive UI, location search, geolocation, 5-day forecast, Leaflet maps, YouTube gallery |
| Backend | RESTful API routes, Turso CRUD, Zod validation, 5-format data export (JSON/CSV/XML/PDF/MD) |
| Stack | Next.js 14 · TypeScript · Tailwind CSS · Turso (libsql) · React Leaflet · jsPDF |
cd weather-app
npm install
# Create .env.local with your API keys:
# OPENWEATHER_API_KEY=your_key
# YOUTUBE_API_KEY=your_key
# TURSO_DATABASE_URL=your_turso_database_url
# TURSO_AUTH_TOKEN=your_turso_auth_token
npm run dev
# Open http://localhost:3000Location: weather-forecast/
An end-to-end data science pipeline analyzing 142,000+ global weather records across 257 cities over two years. Implements the full Advanced Assessment including all five unique analyses.
| Area | Details |
|---|---|
| Dataset | GlobalWeatherRepository.csv — 142,093 rows, 41 columns, 2024–2026 |
| EDA | Distribution plots, wind rose, correlation heatmap, IsolationForest anomaly detection |
| Spatial | Plotly choropleth, Folium bubble map + PM2.5 heatmap, continental comparisons |
| Forecasting | Prophet (daily) + SARIMA(2,1,2)×(1,1,1,12) (monthly) + inverse-MAE ensemble |
| Unique Analyses | Climate patterns, environmental impact, feature importance, spatial, geographical |
cd weather-forecast
python3.11 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
jupyter notebook weather_forecast_analysis.ipynb
# Run all cells: Kernel → Restart & Run AllMIT License — see LICENSE