CoinAnalyst Corp. announced the release of the new AI-based big data analytics platform's backend. Many months in the making, the development team has completed an extensive overhaul of the CoinAnalyst Insights backend architecture.

The entire system behavior has been reengineered from the ground up to provide a fully distributed architecture. The new messaging bus and database storage provide an effective way to now retain far more documents and scale horizontally, increasing storage capabilities from gigabytes to petabytes. Also noteworthy is the fact that the number of messages has been significantly reduced and the lightweight bus has lowered the load on the database, enabling an increase in the overall throughput of the system.

The new plug-in architecture allows developers to add extra steps to the document processing flow and mix different environments for the most suitable tool handling a task. Previously, the CoinAnalyst development team could only work with Java-based applications. Through the universal message bus, they can now include tasks in Python or C++ to gain better performance or features that are not available in the form of Java packages.

This gives the team access to a wide range of AI solutions developed by the Python and C++ communities. This extensible architecture also makes it easier to implement support for new data types and incorporate them into the system with ease. In addition, index-level deduplication ensures that there is only one version of a document associated with a URL in the index.

This greatly improves storage utilization, as the data stored or transferred is greatly reduced. CoinAnalyst's Rest API has been updated and improved to feed both institutions and exchanges with data. The platform can provide structured sentiment values from the text in news and comments.

The social buzz values can be used for trading strategies and quant algos.