HOPEGLOW

Inspiration

Many individuals and community organizations — such as low-income families, patients, orphans, or small charities — are in need of everyday support and donated items. At the same time, countless people are willing to give, but lack a convenient and trusted way to connect with those in need.
More importantly, a person may be a donor one day and a recipient the next — generosity is not one-directional.

HopeGlow is a web-based donation platform that allows users to donate and receive items based on proximity and need. Whether you're offering support or seeking it, HopeGlow provides a fast, simple, and secure space to connect, share, and uplift your local community.

Frontend: Designed using HTML, CSS, and JavaScript, with a responsive UI that supports key features like login, item listing, donation history, and request tracking.
Backend & Storage: Powered by Firebase – including Firebase Authentication for secure user login, Firestore for real-time database management, and Firebase Storage for hosting item images.
Image Recognition with TensorFlow: Integrated a TensorFlow model to recognize and label donated items based on uploaded images, helping to auto-fill item information and improve search accuracy.
Location-Based Matching: Used the Google Maps API to geocode pickup addresses and sort search results by distance, allowing users to find nearby donations more efficiently.

I initially planned to use the OpenAI API to auto-generate item descriptions, but ran into integration issues and pivoted to a TensorFlow-based solution instead.

I developed HopeGlow from the ground up — including user authentication, donation listing, item request, and history tracking — using Firebase for backend services.
I used the Google Maps API to geocode addresses and sort donation results by proximity, helping users find items near them more efficiently.