Andrew Castro | Travel API Project

Travel Suitability API

Built with FastAPI, asynchronous HTTP requests, and scored weather recommendations | November 2025

Author: Andrew Castro

This project is a REST API built with FastAPI and Python that returns travel suitability recommendations for user-supplied cities. It resolves ambiguous locations by combining city input with optional country and state filters, then fetches live geocoding and forecast data from Open-Meteo to generate a normalized trip score and verdict.

The service uses Pydantic models for typed response contracts, asynchronous external requests for responsive API behavior, and production-style deployment settings such as CORS configuration and cloud hosting on Render.

Project Summary

Built a FastAPI REST API that accepts city, country, and optional state input to return travel suitability recommendations from live weather data.

Integrated Open-Meteo geocoding and forecast endpoints using asynchronous HTTPX requests to resolve locations and retrieve daily forecast values.

Handled ambiguous city names by filtering geocoding candidates against optional country and state parameters before selecting a final destination match.

Implemented typed Pydantic response models for predictable JSON output, validation, and cleaner API contracts.

Designed a custom recommendation algorithm that scores temperature, wind speed, and rainfall on a normalized 0-100 scale and returns a final trip verdict.

Tech Stack & Methodology

Built with FastAPI and Python for REST endpoint development and request handling.

Used HTTPX for asynchronous external API calls to geocoding and forecast services.

Structured responses with Pydantic models for typed validation and consistent output formatting.

Hosted on Render with CORS configuration and production-style API deployment settings.

Key Engineering Challenges Solved:

Location Disambiguation: The API evaluates multiple candidate cities and filters them using optional country and state input to avoid incorrect matches.

Failure Handling: Returns clear HTTP errors for missing cities, invalid matches, and unavailable upstream data.

Data Transformation: Converts geocoding and daily weather responses into a compact, readable recommendation payload for frontend consumption.

Live API Playground

Interact with the live API below. The form sends a request to the hosted FastAPI backend, which resolves the destination, fetches current forecast inputs from Open-Meteo, calculates a recommendation score, and returns a structured result.

The API is hosted on Render's free tier, so the first request may take a few moments to wake up.

Conclusion

This project demonstrates backend-focused software engineering through API design, external service integration, typed response modeling, and deployment to a live cloud environment. By separating the recommendation logic into a standalone REST service, the application can support web clients and other frontend consumers without coupling business logic to the interface layer. It highlights maintainable Python development, practical API error handling, and real-world integration with third-party data providers.