The Problem
Local marketplaces like Facebook Marketplace, Craigslist, and others have become the default way people buy and sell secondhand goods, but the experience hasn't meaningfully improved in over a decade. Sellers are forced to guess at fair prices, often underpricing valuable items or over-pricing common ones, which leads to listings sitting unsold for weeks. Buyers waste hours scrolling through poorly-described items, haggling over cash prices, and coordinating awkward in-person meetups. On top of that, most transactions still revolve around cash, which excludes anyone who doesn't want to deal with the friction, the safety concerns, or the mental overhead of constantly negotiating dollar amounts for low-value items.
Another problem that lies behind almost every modern app is their analytics. Most developers default to Firebase Analytics because it ships free with the rest of Google's stack, but Firebase was designed as a general-purpose tool and can struggle for specific cases. Event data takes up to 24 hours to appear in dashboards, custom event definitions are rigid and require redeployment to change, debugging individual user sessions is nearly impossible, and the UI buries the few insights you actually need under hundreds of metrics you don't.
Project Overview
Throwly is an AI powered marketplace application meant to create an environment for users so they do not have to worry about managing pricing, funds, complicated communication, and difficult deliveries. It reimagines the local marketplace by using points instead of cash, and AI-powered pricing instead of guesswork. Rather than haggling, users trade items using points that the app awards based on the estimated value of what they list. When a seller uploads photos, the pipeline sends the listing through a AI model that has been benchmarked against real eBay sale data and it returns a fair point value, a confidence score, and a category. This removes the negotiation step entirely and means every trade feels balanced, regardless of whether someone is offloading a couch or a coffee maker. The mobile app is built in React Native with Expo, the backend runs on Firebase Functions, and the AI pricing layer is evaluated continuously against a labeled dataset of real listings to keep its estimates accurate.
Distri is meant to be an alternate analytic dashboard system. Through expandability and simplicity, the goal of Distri is to allow the user to manage their applications analytics without much stress and setup. Distri runs on an independent website that is connected to an external application via NPM package. Users are introduced with a package download and setup instructions that are easy to follow. Once connected, users do not have to worry about much else as all of the data is sent and managed in Firebase before being displayed on the website. Distri creates a unique experience by "gamifying" the usual analyti dash flow. We took inspiration from the very common "season" structure in modern video games to track user activity over a set period of time. Once seasons are over, the top user may be rewarded with something of the project owners choice. This process is safe and simple with automatic season renewing, automatic email prize claiming, and a flagging system to catch event spam. Distri uses a very simple flow where users simply login/signup -> create project -> attach NPM package -> view analytics to keep effort as low as possible.