PolicyTool: Determining Country Happiness from Social Media Data

Won 2nd award in the in the Physical Science and Engineering Category of the 2022 Synopsys Championship (Santa Clara county division). Understanding the dynamics of citizen happiness would be massively beneficial for policymakers, given that their decisions can correct downward trends in happiness. Doing so is as important as ever in today's society, but existing methods such as the one of the World Happiness Report are too inefficient. My solution, PolicyTool, decreases turnaround times to mere seconds while being extremely cost efficent and just as accurate as existing methods by harnessing the power of widely accessible data on social media platforms. It is an entire end-to-end data ingestion pipeline that processes Twitter data and performs data analytics to evaluate the overall happiness of a country, with handcrafted individual software components for job dispatch, distributed computing, data/buffer replication, and advanced SQL queries.

Fern Programming Language

Functional programming has brought monumental gains to the developer experience, but it still has pitfalls, most notably performance when frequently reading and writing to random-access data structures. Fern is a general purpose programming language I created, transpiled to C, that presents the performance of random-access data structures while preserving the developer experience of FP by providing a persistent interface to the data structure. It is up to 20x faster than other FP languages, including Haskell and Elixir.

Worm DSL (Domain Specific Language/Active Code Generator)

Worm generates code for the Phoenix framework, a framework for building backend applications that run on Elixir. It keeps the benefits of working popular backend frameworks (reliability, performance, support, etc.) while massively reducing the downside, namely boilerplate code, by generating database schemas and server routes with a powerful syntax. It serves my co-founded startup ezML by speeding up development and promoting clean, efficient code.

Chess Enabled

Enables the most people to have the best chess experience through reliability, fairness, and accessibility. Features an elo-based matchmaking system for maximum fairness and voice input functionality using the Alan AI SDK to help those that are imparied or simply want more convenience.

Edge

Allows for the containerized remote execution of code in a high-throughput P2P computational pool, perfect for running on the edge. Client side interactions are done with a Python CLI, along with raw network sockets for P2P and FTP, and a load balancer built in Nest.js.