Changelog
Track our journey as we build the future of data visualization for trading and investing.
Updating Dependencies & Tech Stack
Let's get with the times, once and for all.
In this release, we focused on modernizing our tech stack and updating key dependencies to ensure Rthmn remains at the cutting edge of web development best practices. We made significant changes to our backend infrastructure, frontend frameworks, and data storage solutions.
Key updates include:
- Switched to Bun and Elysia.js for our production server - We migrated our backend to the lightning-fast Bun runtime and adopted the Elysia.js framework for building high-performance, scalable APIs. This move allows us to leverage the latest advancements and deliver even faster response times to our users. Fun fact: Yes, we're building trading algorithms in TypeScript.
- Upgraded to Next.js 14+ for our client app - To provide the best possible user experience, we upgraded our frontend to the latest version of Next.js. This update brings performance enhancements, improved developer experience, and access to the latest features and optimizations offered by the framework.
- Embraced Supabase and moved off MongoDB - As part of our commitment to using the best tools for the job, we made the decision to migrate our database from MongoDB to Supabase. Supabase's PostgreSQL-based solution offers better scalability, more robust querying capabilities, and seamless integration with our existing tech stack. Also they have been releasing very cool realtime features we plan to use.
- Implemented Tanstack Query for efficient data fetching - To optimize our data fetching layer, we adopted Tanstack Query (formerly known as React Query). This powerful library simplifies the management of asynchronous data, providing built-in caching, background updates, and error handling. The result is a more responsive and resilient application.
- Progress on user authentication flow for paid access - In preparation for offering paid access to advanced Rthmn features, we began working on a secure and user-friendly authentication system. This will allow us to seamlessly onboard new users and provide a smooth experience for accessing premium functionality.
Creating Calculation Core
How can we translate this into something... Real?
This release introduced the heart of Rthmn's capabilities that paved the way for fully embracing our new method in pattern recognition. This release was the development of our data transformation logic that we designed to convert raw market data into binary sequences of positive and negative values. This allowed us to identify and track market patterns at various vector scales.
While Rthmn is not a time-based algorithm, it builds upon the valuable insights gained from time-based trading strategies. These strategies have long recognized the significance of patterns in market behavior. However, the missing piece has always been the realization that the shape and structure of these patterns hold even greater importance than their timing alone.
By removing time from the equation, Rthmn naturally uncovers more consistent and identifiable patterns. This shift in perspective also allowed made the development of building a robust systems far more intuitive as our source of truth is that the market is always producing recurring market structures. As far as we're concerned, all we need to do is package and containerize the patterns, and understand how they start and how they end.
In this release to our work we shipped:
- A robust historical data fetching + real time system
- Calculation and processing layer after data is being fetched
- Storing the calculated data in a scaleable way for further processing, used for backtesting and future planned features
Starting Building A Lab
We need a place to experiment, test, and learn. We need a lab.
Building upon the findings and capabilities introduced in v0.02, this release focused on further optimizing and refining our core algorithms. We made significant strides in improving the efficiency and accuracy of our real time calculation flows and what will later turn into pattern detection and matching processes. Think signals, (eventually).
In this release, we also started a local-only lab environment we aim to continue to maintain for our internal purposes where we perform greater historical tests and use this platform to act as a staging ground for shipping new features to rthmn.com
Key enhancements included:
- Established a robust testing environment - To ensure the reliability and performance of our pattern recognition system, we built a comprehensive testing environment that allowed us to validate our algorithms against diverse market conditions. This rigorous testing process helped us identify and resolve potential issues before they could impact our users.
- Implemented a continuous integration and deployment (CI/CD) pipeline - We streamlined our development process by adopting a CI/CD approach, enabling us to deliver updates and enhancements to the production system rapidly and with minimal downtime. This agile workflow allowed us to respond quickly to user feedback and market changes.
By continuously refining our algorithms and infrastructure, we are ensuring that Rthmn will remain at the forefront of pattern-based trading innovation. Our unwavering commitment to performance, accuracy, and user experience are setting the stage for the transformative features that will define the future releases of our platform.
Building Boxes
Every story has a beginning. Ours began with a box.
The initial idea for rthmn was originated in 2019. Since then, a ton of progress has been made towards systematizing it and applying it for more universal use cases.
Our first milestone was to build the core logic for translating market data and building a maintainable engine where we could understand the steps involved. This involved creating a custom rendering pipeline optimized for displaying complex financial data in an intuitive and interactive way.
We started by designing the data structures and algorithms needed to efficiently process large amounts of historical and real-time market data and always stay in sync. Throughout this process, we stumbled upon even more novel ideas that soon began to become stable methods we even rely upon today to plan features.
Scalability was another critical consideration. We knew that as rthmn grew in popularity, we would need to handle a rapidly increasing number of users and data points. To address this, we designed our backend infrastructure to be highly modular and horizontally scalable.
In addition to the database, Rthmn's architecture is designed for scalability across all components. The application is built using modern, efficient technologies and follows best practices for performance optimization. This includes techniques such as lazy loading, caching, and asynchronous processing to ensure a smooth user experience even under heavy load.
As the platform continues to grow, we're continuing to regularly monitor and optimizes the infrastructure to maintain optimal performance.
This is only just the beginning.