Creating and Combining Alpha Streams from Big Data

Armando gives a short introduction to the event and briefly examines what is Big Data in Finance

Big Data through digital transformation has the promise to transform every industry and quantitative investing is no exception. Automated sentiment analytics on large volume and high velocity news streams is here, but what else is available to be exploited for investment research and trading opportunity detection? Where is the big data market heading in terms of content and applications that will be relevant to quantitative based investment research and new strategy formulation? In this short presentation, Nicolas explores new potential big data use cases, how can big data be made available and exploited, and what are the main data and analytics technology patterns to practically overcome volume, variety, veracity and velocity issues.

Market response to sentiment derived from news events is not linear. Indeed, strategies based purely on positive versus negative sentiment are generally sub-optimal - either because the signal is too “noisy”, has a high turnover, or both. In this presentation, a thematic approach to sentiment analytics is proposed. Specifically, Peter shows how to construct thematic alpha streams, find out which themes are key drivers of performance, and discusses how individual thematic alphas can be combined to create stronger portfolio-level alphas.

At its essence, quantitative investing relies on the identification of signals to generate alpha. One of the main criticisms of this approach is that the signals created often rely on company data and can therefore lack human insight. In modest terms, a machine simply cannot read and react to the news in the same way that an analyst can. With recent developments in big data, we are starting to see a shift in this paradigm. In this presentation, Russell is looking at how big data concepts, specifically RavenPack News Sentiment indicators, can be combined with traditional factor strategies to provide more intelligent factor signals.

In this brief session, Steve introduced three methods to interface RavenPack News Analytics with MATLAB. He discusses how you can incorporate earnings announcements and news sentiment, for example layoff data, to inform historical analysis and back-testing and your intraday trading strategies in MATLAB

Central bank communications are an essential driver of today's global markets. With nominal interest rates constrained by the zero bound, central bankers are guiding investors' expectations with their own words. While investors' attention often focuses on only a few specific words, such as the ECB's "whatever it takes" and the Fed's "patience" for example, NLP techniques pay attention to the overall message carried by Mr. Draghi. ECB press conferences are particularly important among the central bank's communications because they are delivered live and include unprepared Q&A. This allows us to track each sentence within each topic and measure their immediate market impact. This presentation highlights, in particular, that NLP now belongs in the standard toolbox of every professional quant analyst.