Note: I used Pythonanywhere and this is the result (https://csusanto.pythonanywhere.com/)
This project is a small Flask web application that helps users find the nearest MBTA stop to a place name in the Greater Boston area and also some weather data. When the user types a location (for example, “Boston Common” or “Harvard Square”), the app:
- Uses the Mapbox geocoding API to convert the place name into latitude and longitude.
- Uses the MBTA v3 API to find the closest transit stop and reports whether it is wheelchair accessible.
- “Wow!” features include the use of the OpenWeatherMap API to show the current temperature and weather description near that location.
# Development Process
When working on all types of projects, I like to break things down into steps so that everything is coherent and able to be backtracked. The most challening part of this project was using the pythonanywhere since I wasn't familiar with it. Until it took some time for me to realize that its just a terminal that runs my code and that this website had the ability to track the count of users who use this code. (https://csusanto.pythonanywhere.com/)
Testing the code definitely made it easier to understand which is a big takeaway from me.
One of the trickier parts was understanding the nested JSON structures from the APIs. I used `print` / `pprint` and sample responses to find the correct paths like `["features"][0]["geometry"]["coordinates"]` for Mapbox and `["data"][0]["attributes"]["name"]` for MBTA.
# Teamwork & Work Division.
Since I worked alone I told myself to create a step by step process and to make it foundationally strong so that I can rely on finishing one task before moving on to the next
# Learning & Use of AI Tools
Chatgpt helped me alot to help speed up clarifying parts of the process that I had no prior experience in.
It helped with suggesting how to build URLs and handle errors.
Walking through deployment steps on PythonAnywhere, especially WSGI configuration and `.env` issues.
One limitation of Chatgpt was that it sometimes didn't understand the goal of the project so I took alot of time to add details to the descriptions to make sure that we were on the same page.
Looking back, I wish I had used AI a bit later in the workflow—first trying things manually, then using it to confirm or improve my approach—because that would force me to think through more of the errors on my own.
I've realized how easy I have it as someone who lives in the 20th century to have all these helpful tools to assist me in making a website and I have a nice computer to help me. I can't imagine people in the early 1990s or 2000s doing this process.