Corporate Competencies

Data cleansing, restructuring and cross-referencing

RavenPack’s relational databases ensure that data – whether stock market tick data, weather figures, economic statistics or unstructured news XML – is consistently time stamped, formatted, cleansed and structured to meet the needs of clients demanding the highest level of data reliability and consistency.

Collection of massive amounts of news via direct real-time feeds

RavenPack’s data warehouses connect to numerous high-frequency news feeds (Dow Jones, Reuters, Thomson Financial, etc.) while leveraging supercomputing power and massive storage capabilities. Decades of historical news on the multi-terabyte scale are standard implementations of the company’s infrastructure.

Parallel databasing and analysis

Just as RavenPack’s servers are able to provide managed storage for dozens of high frequency data feeds, they are capable of parallel, real-time analysis of such feeds. Analysis consists of sentiment scoring, data normalization, metadata application (subject, genre, facts, and sentiment tags), content structuring, relationship identification and others. These processes run in millisecond time on all events for all feeds in parallel.

Data classification and tagging

RavenPack has accumulated significant expertise in classifying and tagging data streams, particularly those related to financial news. Dozens of automated classifiers identify whether stories are positive, negative or neutral using varied linguistic techniques developed by RavenPack’s in-house experts. Additionally, classifiers identify, extract and structure story substance ranging from aboutness, genre, key actors, locations, impact, relevance, facts and figures, and others.

High availability data retrieval systems

Subscribed users leverage the company’s data-mining capabilities that allow for instant retrieval of time-series data and analysis via our API. RavenPack’s proprietary, hosted data structures enable access to massive, high-maintenance databases through simple, zero-cost web interfaces.

Customized news sentiment and volume indicators based on client's area of interest

Leveraging the company’s data infrastructure against extensive news archives from diverse sources, RavenPack is able to construct news sentiment and volume indicators for varied markets, asset classes, financial themes or individual companies. Extensive linguistic experience empowers classification methodologies that are both robust and modular.

For more information on RavenPack, email info@ravenpack.com.