Overview

Sigmai’s data works well for any company that wants a thorough and automated understanding of company-specific business events published in the media.  For example; on any given day Sigmai’s data can tell you that: Apple released their financial results, settled a lawsuit, a media article mentioned the CEO, and any number of over 400 other specific events.

With more than 100 million events identified, researchers and data scientists will use our feed, or a local instance of our software and combine it with other data sets for unique insights. Their analysis can yield strategies for trading, risk management, customer identification, credit ratings.

More specifically, we’ve found that equity research firms and hedge funds have an interest in Sigmai’s products.  Firms that research and monitor the reputation of corporations and CEO’s will also find the data valuable.

 

 

Case Studies

  • Combining the events with share price movements and analyzing this relationship has already yielded insights that have been used to build medium-term trading strategies.
  • We’re currently trialling trading strategies that act on the relationship between the rare and random events that we detect and the share price volatility that follows.
  • Analysis has been done to establish CEO media exposure profiles, and to compare these exposure profiles across companies within a sector.  The resulting insights can be used for investor and public relations departments to develop and monitor publicity strategies.

Additional Information

The software isn’t perfect, but we’ve got a proprietary PDQ system for that:

A specific example is that our natural language processing algorithms will sometimes identify content provider companies as related to the business event that they are publishing.  An example is Sinclair Media. To correct this, we divert the business events associated with Sinclair Media through a more rigorous quality control process. This quality control process features our proprietary PDQ system that keeps the accuracy of our data high by using efficient and targeted human input.

 

There are two sources of sampling bias in Sigmai’s data.  The first is that, despite the data being drawn daily from thousands of news sources from around the world, the news sources must be in English.

The second source of bias is that of the nearly 4000 companies covered by our service; most are American and trade publicly.  We are continually improving our coverage by expanding our library of organisations as clients request the inclusion of new entities.

Sigmai’s data faces no issues regarding data privacy.  All the information we collect and analyse is publicly available. Also, the data sets we build do not contain sensitive information relating to any individuals.

Our core technology is built to understand language. The creation of Sigmai’s systems was a truly monumental and complicated task.  We took on the challenge at a level of detail that no other company in the world chose to tackle. Our approach took a team of academics, statisticians, data scientists and linguists over 50-man years of effort to develop and is now the leading solution of its kind.