SingleStore Hackathon 2022: Here Are the Big Winners


Dave Martin

Director, Developer Relations

SingleStore Hackathon 2022: Here Are the Big Winners

The 2022 SingleStore Hackathon didn’t disappoint. Across all five categories, our team was blown away by the submissions we received. 

After the Hackathon closed and we had the opportunity to dive deeper into submissions, we live streamed our award ceremony to announce the winners. Here’s who came out on top of each category, and took home their share of prizes. 

Category 1: Multi-Model

Requirements: Build a multi-model application that supports JSON, geospatial, time-series or full-text search.

Winner: Animal Explorer, a simple Node.js web application with SingleStoreDB and AWS that creates interactive animal discovery through map and geospatial data.

Description: Animal Explorer is a web application that allows users to know where animals are all over the world, in real time. According to creator Huan Ho, the app offers two key functionalities:

  1. An interactive map where users can draw any area
  2. A search functionality that retrieves the updated animal list stored in SingleStoreDB.

To build the application, the Animal Explorer team used Node.js and Express.js along with HTML, CSS. JavaScript, Bootstrap 5.2 and Leaflet.js to create both the server and user side.

Try it out: Animal Explorer Github Repo

Category 2: Hybrid Transaction/Analytical Processing (HTAP)

Requirements: Build an application that requires a high volume of transactions along with analytical queries on big and/or streaming data. 

Winner: Router Holmes, an application that takes in the real-time stream of data received by internet routers, analyzing diagnostics and displaying that data on a map. 

Description: Creators Vaishakh Sreekanth Menon and Dhushyanth Sundararajan drew inspiration from IoT companies providing internet-related services: “We realized that such  companies must be getting data at a high frequency, and we also came across several use cases for such data…”

Router Holmes’ creators go on to explain that they felt they could easily leverage SingleStoreDB features for real-time analytics to receive and analyze the internet data. In addition to using SingleStoreDB, this Hackathon entry was built using Kafka, MySQL and React.

Try it out: Router Holmes Github Repo

Category 3: Wasm/WebAssembly

Requirements: Implement any calculation or function that does not exist in SingleStoreDB today by writing a new Wasm UDF. Alternatively, create a spaceship strategy that competes in the Wasm Space Program. For either, write your code in Rust, C++ or C, and execute it in SingleStoreDB.

Winner: Drone Surveillance, an AI-based path optimization within SingleStoreDB.

Description: According to creator Nakul Havelia, inspiration for the project came from the rising use of drones in industries like construction, oil & gas, disaster management, last-mile delivery, healthcare and more. The project includes a Wasm function that calculates the optimal path for several drones.

First, it clusters the geo-coordinates, then runs the Simulated Annealing (SA) to solve traveling salesman issues. The algorithm for path optimization was built using Rust and made available in SingleStoreDB in the form of a Wasm UDF. Additionally, Nakul used HTML and JavaScript to simulate the drones.

Try it out: Drone Surveillance Github Repo Part 1; Part 2; and Part 3

Category 4: Machine Learning (ML)/ Artificial Intelligence (AI) 

Requirements: Train and/or execute a ML or AI model entirely within SingleStoreDB.

Winner: Movie Recommender, an application that allows you to spend time watching — not searching — for movies.

Description: For creators Jason DSouza and Siddhant Mohile, inspiration for the app stemmed from a joint desire to create a movie recommender that suggested titles they each enjoy. Jason and Siddhant started by uploading data into SingleStoreDB, then created a Streamlit website where users can enter their gender, age and genre preferences — information that was then queried with SQL statements.

Movie Recommender uses an ALS algorithm to generate a number (factors) for each user and movie. The numbers range between 0.5 and 1. When multiplying the factor of a user with the factor of a movie, a total number is calculated. The higher that number is, the greater chance a specific movie is recommended for that specific user. By using the correct SQL statements, the Movie Recommender application can present the top 10 movies viewers would watch based on their gender, age and genre preference.

Try it out: Movie Recommender Github Repo

Category 5: Database Migration 

Requirements: Migrate any application to SingleStoreDB.

Winner: Airbyte SingleStore Connector, a destination connector for the widely used, open-source ELT platform.

Description: Creator Mohamed Magdy describes AirByte as an easy-to-manage and deploy ELT tool. But, the current sources and destinations AirByte ships don’t include SingleStore — which encouraged Mohamed to build one for the Hackathon.

The SingleStore destination is designed to be used with any Airbyte source, writing imported data from the original source directly to SingleStoreDB. 

Try it out: Airbyte SingleStore Connector Github Repo 

dont-miss-out-on-the-next-single-store-hackathonDon’t Miss Out on the Next SingleStore Hackathon! 

We’ve got another incredible Hackathon coming your way this winter — stay tuned for updates by following us on Twitter at @SingleStoreDB and @SingleStoreDevs.

Want to start prepping? Try SingleStoreDB free today.