2026

FlightNu

A flight price intelligence platform that tracks fares over time, runs statistical forecasting, and tells you exactly when to buy.

Solo Full-Stack Engineer
overview

FlightNu is a full-stack flight price analytics platform built from scratch: a Bloomberg-style dark-mode dashboard that continuously monitors flight routes, runs statistical analysis on historical fare data, and generates buy/wait/monitor recommendations backed by real math. It features a synthetic price simulator with route-distance modeling and seasonality curves, a Python analytics engine computing moving averages, z-scores, and linear regression forecasts, an APScheduler background worker that ingests price observations every 4 hours with no browser required, and a Next.js frontend with glassmorphism cards, live Recharts price history overlays, and per-route prediction confidence scores.

FlightNu

The goals

  • Turn raw flight price observations into actionable buy/wait signals backed by statistical forecasting
  • Build a persistent monitoring system that runs continuously in the background, independent of any user session
  • Design a fintech-grade analytics dashboard that makes complex price intelligence feel immediate and readable

What I did

  • Built a Python analytics engine from scratch (SMA, EWMA, z-score anomaly detection, linear regression forecasting, 95% confidence intervals) plus recommendation logic with explicit statistical thresholds, every BUY NOW signal ships with a plain-English explanation
  • Designed a synthetic price simulator modeling route-distance base fares, day-of-week and hour-of-day seasonality, days-to-departure urgency curves, and stochastic volatility spikes, swappable for a live pricing API in one file
  • Shipped a FastAPI and SQLAlchemy backend with a full PostgreSQL schema and an APScheduler worker that ingests prices and recomputes analytics every 4 hours server-side, deployed on Render with zero browser session required
  • Built a responsive Next.js dashboard with live price history charts, SMA/EWMA and volatility overlays, price histograms, seasonality analysis, and a tracker creation modal with per-route Discord and email alerting

Live impact

4hrs
Price observation cadence, fully automated server-side
120+
Historical observations seeded per tracker on creation
3
Statistical models running per tracker (SMA, EWMA, linear regression)
95%
Avg prediction confidence on seeded demo routes
<2s
API response time for full dashboard summary with analytics
0
Browser sessions required to keep tracking running

Tools & skills

Next.js 16TypeScriptTailwindCSSRechartsFastAPISQLAlchemyPostgreSQLAPSchedulerNumPyPydanticRender

The results

Live at flightnu-web.onrender.com with persistent background price tracking across all active routes
Full statistical analytics pipeline: moving averages, z-score anomaly detection, linear regression forecasting, confidence intervals, and seasonality analysis per route
Recommendation engine with explicit mathematical thresholds: every BUY NOW signal is explainable
Solo full-stack build shipped end to end: analytics engine, price simulator, REST API, background worker, and fintech-grade UI
FlightNu screenshot 2
FlightNu screenshot 3
FlightNu screenshot 4