Understanding price movements in financial markets is all about sentiment. On the face of it that might be an odd statement to make because surely prices defend on more fundamental factors than mere sentiment?
That’s true at one level. Whether a company is making a profit, how much revenue it is generating, how much free cash flow it has, what its earning per share are, winning a big new order, making new partnerships, take-off bids, and so on, are all important considerations to be assessed by investors when making buying and selling decision on stocks and shares.
When markets are functioning as they should, much of that information is available to all as required by law when it comes to public markets, so all investors should be on a roughly level playing field. Although professional analysts at the large investment management companies and banks may have an advantage because of the direct line their analysts will often have to the companies they cover.
How investors react to company announcements will influence prices, often dramatically, which is where sentiment analysis comes in.
If a company increases its profits you might expect its share price to rise, but if earnings per share don’t meet market expectations as set by analysts across the industry, the share price could still fall. Anticipating the reactions of professional investors and the ordinary retail investor is critical for traders to get ahead in the market. It is why Wall Street has been taking Big Data and sentiment analysis so seriously, with computer-driven “quant trading” having been around for a few years now.
And it is not just the headline-grabbing news flow that matters. In the many weeks, days and hours between quarterly market-moving earnings reports, investors and traders are making decisions all the time, often by taking into account both the fundamental analysis from business performance and the technical analysis of price movements.
All of the millions of decisions made by traders, taken together, creates the price movements reflected in technical data and the profit and loss showing up in individual investor portfolios. Sentient analysis is a way to get underneath the macro picture by making sense of the millions – if not billions – of pieces of information that are generated by investors that reflects their opinions and stance towards different companies, sectors and geographical regions that may be considering investing in or getting out of.
Sentiment Analysis for All
Today, for the first time, the attempt to analyze this mass of sentiment data represents has moved beyond the deep-pocketed investment banks and is now, with the help of blockchain, Big Data and artificial intelligence (AI), accessible by private investors.
Global data provider Thomson Reuters recently added to its MarketPsych Indices a bitcoin sentiment data feed that employs AI to analyze 400 data sources to provide subscribers with predictions upon which they can base their decisions.
By capturing what market participants are posting on social media, the news and analysis in the financial and wider mainstream media, and even the contributions of attendees at industry conferences if there’s a consumable live audio feed, all this data can be plugged into a sentiment analysis engine powered by AI.
There are a host of blockchain-based projects focusing on trading solutions using sentiment analysis. Capitalise has a beta running that makes it easy to set up triggers to execute trades, powered by real-time sentiment analysis data. To keep down costs Capitalise has partnered with Senno, a blockchain sentiment analysis platform with an open API.
Senno is interesting because its offering is aimed at providing a lower-cost solution that can be employed by smaller companies and individuals. The software development kit (SDK) is downloadable for free and the application programming interface (API) is public, so both corporates and private citizens can build applications with it to connect real-time sentiment analysis and business intelligence analytics to their ecosystem.
Focusing on algorithmic trading platforms and advertising/marketing firms, Senno has cleverly positioned its product so that it has a relatively shallow learning curve for adoption and much more manageable integration expenses. Behind the software, Senno deploys a distributed hardware solution that allows it to deliver a relatively low-cost solution to its clients.
There are many other players in the field. Worth a mention is Santiment. Its Sanbase platform measures crowd moods expressed in the “sentiment wave”, where market participant attitudes can swing quickly between excitement and despair (particularly prevalent in crypto markets!) and how to apply insights to make informed investment decisions.
Another promising project is Token AI with its proprietary Juliet sentiment analysis engine. It offers three service levels: “bottom-up” individual coin analysis, portfolio rebalancing and, if you are a total novice, the token basket generator. The generator works a bit like a robo-adviser –you provide the money, answer some questions about your risk profile and provide other parameters and a basket of tokens is selected for you.
Understanding language, as you would expect, is a crucial part of sentiment analysis. This is where natural language processing science and machine learning comes in. The accuracy of interpreting the social media mentions of a cryptocurrency depends on understanding language usage. The difficulty in doing this is most clearly appreciated when you consider the interpretation of irony and sarcasm, in which the written word is not what it seems. “Verge is down 37%. Thanks for the pump John McAfee, that’s really good!”. That might seem like an endorsement of McAfee promoting the Verge (XVG) coin when it is in fact a sarcastic comment where the word “good” should be understood as “bad”.
Getting to Really Know Your Customers in a Cost-Effective Way
Sentiment analysis is useful for most industries, not just financial markets.
Customer sentiment analysis is an extremely relevant use-case, in which companies can get to grips with understanding and assessing much more deeply the likes and dislikes of their consumers. Sentiment analysis in a consumer setting furnishes companies with a tool to quickly react to customer needs and address problems as they arise and before they multiply.
Being able to produce accurate predictions depends on analyzing in real-time a huge amount of data, which is where the decentralized characteristics of blockchain and its ability to harness disparately located underutilized computing power comes into its own. This makes sentiment analysis much easier to come by for small and medium-sized companies or even for bigger companies that don’t have access to the requisite data sources.
Senno appears to be the first sentiment analysis platform with an open API available for third parties to use. It is built on the NEO blockchain platform and because of that, it can make use of the platform’s digital identity technology to determine and assign trusted data sources, pulling in feeds from social media, forums and messaging apps for a real-time crowd-sourced service.
Another interesting project with a slightly different approach is Sether. It, too, is using the big three of blockchain, Big Data, and AI to engineer its platform. Sether is deploying sentiment analysis to help marketers, influencers, and entrepreneurs interact with their audiences more effectively by, for example, creating more engaging social media campaigns.
Emphasising social media trend intelligence is Bottlenose, although its flagship product – Nerve Center – is not a blockchain platform. It has, however, gained a following because of its effective use of data mining and pattern recognition to provide solutions ranging from consumer insight to risk intelligence.
As we have seen then, sentiment analysis is already being applied profitably and this trend is likely to continue to grow exponentially.
Big Data can increase operating margins by as much as 60%, according to a report by McKinsey Global Institute. Much of the benefit of the technology has up until now been confined to large corporations. That’s all about to change, with new nimble players bringing affordable solutions to smaller companies and individuals alike.
How else can blockchain technology be used to help gather and interpret data for sentiment analysis? Let us know what you think in the comments below.
Images courtesy of AdobeStock
