WSJ: How computers trawl a sea of data for stock picks



Image courtesy of Flickr – provided under Creative Commons License 2.0 here


How computers trawl a sea of data for stock picks

“NEW YORK—In SoHo offices where robots occasionally ply the hallways, dozens of Ph.D. scientists with degrees in fields like astrophysics, immunology and linguistics huddle every day around computer screens that show billions of dollars zapping around the world.

Their goal: to give their secretive hedge-fund firm a leg up in investing the $24 billion it has under management. Scientists at the firm, Two Sigma Investments LLC, program its machines to cull torrents of information from sources like newswires, earnings reports, weather bulletins and Twitter….”

Full story here: – 


QMG Opinion

This is a great article on the use of big data and data mining to get a better view on the stock market. Using data to identify signals and inflection points can be a great way to produce trade suggestions. Examples of consumer sentiment towards products turning from positive to negative are common place in the world of sentiment analysis and this could give credence to a decision to short or buy stock.

The great thing is that software and machines are getting smarter and more readily available to the masses. However some factors should be considered. There have been quite a few attempts to trade on sentiment analysis alone and whilst a lot have been successful there are still ways in which sentiment can be manipulated and gamed (eg employees of Company A contributing to a lot of positive tweets for Company A). This is something that should be within the consideration bucket for any good analysis software and also for data analysts. No matter what, companies will likely try to find ways to positively affect their performance on scorecards of this nature.

So how can you increase the reliability of analysis like this? By ensuring the framework upon which you make those decisions is rock solid and you can do this using factual data. This is data that comes from reliable sources such as government agencies or data collection groups. Coupling that type of information that can be relied upon along with consumer sentiment data and additional stock picking analysis (fundamental, technical etc), means that inherent problems that may occur within each system can be hedged to produce the best alpha capture hybrid tool possible.

The nature of the facts is another ball-game all in itself and there are certainly quite a diverse range of providers out there, from government agencies to private companies (large and small). The availability and reliability of such data can be called into question as well. The availability becomes a problem when only certain industries or types of data can be used and reliability becomes a problem when a dataset hasn’t been used before.

At QMG, we’ve done a lot of work to solve these puzzles. Firstly, we obtain data from national statistical organisations (NSO’s) spanning the globe. Then, we enhance that by using our rigorously tested 20+ year statistical model that can extrapolate earnings data (Sales, Margins) from what drives these figures (Price, Cost, Volume). Further breaking it down to the sub-sector categorisations gives a more focused view that is not obtainable anywhere else.

So how would factual data, pieced together in a system like ours, help clients in their funds management process? Without focus and clarity, investment managers are left trawling through a universe of stocks, and screen based on very broad factors such as which stocks are affected by the ISM manufacturing index numbers, or overall GDP growth in the Materials sector. Instead, QMG quickly and easily screens what clients care about at whatever level that is (region, industry, sector, stock etc) and produces bespoke solutions that become part of the clients day-to-day investment process.

Here’s an example of how the process works:

Step 1 – Clients advise what countries/industries/sub-sectors/stocks/indexes they care about

Step 2 – QMG’s analytical model processes those stocks in relation to our coverage universe and produces e a scorecard highlighting key areas of focus


Step 3 – Results of screening process produce smaller basket of opportunities that can be worked on further via clients other processes

The advantages of running it via a system like this means that alpha capture is quick and reliable. Furthermore, trends can more easily be identified – for example our note on Global Autos (click here) came purely from building a system that notified us of interesting data coming out of automobile related sectors.

Example page from Global Autos Report – April 2015

Screenshot 2015-05-02 21.56.53

Decision making in the financial world has not been made any easier with all the noise out there. QMG can help decision makers can make better choices, and coupled with highly factual information, delivered in a timely manner, the advantages that we can give to those who analyse consumer sentiment are boundless.

For more on how QMG can help you can contact us here

Leave a Reply

Your email address will not be published. Required fields are marked *

× eight = 56