Employing Social Sentiment Data for Investment and Trading: An Introduction (Part 1)

This article is the first part from a short series of articles about investment and trading using SSIX social sentiment data.

As social networks become more and more part of our daily life, an increasing volume of comments and opinions covering a huge variety of topics are shared every second. Considering that social networks are continuously increasing their user base from all generations, they start to be a more and more accurate reflection of our society.

In SSIX project, we are developing methods and technologies that target specific topics of interest in Finance, CRM, Politics or News, in order to translate semantic data into usable and actionable Social Sentiment data.

In this article we are going to focus on the financial area, in particular on how Social Sentiment data – a new alternative data type for finance – can be used for investment and trading in securities, with a special focus on stocks.

Investment and trading professionals need accurate and clean financial data in order to find specific patterns. Such patterns give them indications about how various securities are being perceived by the markets as well as about their possible future evolutions and moves.

Usually, financial data comes from a variety of sources such as companies’ annual and quarterly reports, covering balance sheet data, income statements, cash flows as well as market data such as price and volume. Based upon price and volume, many other technical parameters are build such as MACD, EMA, Bollinger bands etc. Overall, a stock can be described by more than 1,000 parameters. Moreover, for particular investment and trading styles, understanding the market sentiment is also crucial to define the best security selection and execution strategies. In this respect SSIX project provides a set of social sentiment parameters (or X-scores) that can be added and mixed with any other financial parameters for a better decision making process.

However, if trading and investment were to be decided and executed solely on social sentiment, such operations could be highly volatile and risky. Unless one wants to use high risk approaches, we would not advise to solely use sentiment parameters to determine trading and investment patterns, but to use them combined with other financial parameters such as fundamental and technical ones. In this way, the social sentiment influences the overall strategy in a balanced way rather than disproportionately.

Therefore, in order to use the social sentiment data optimally, it isn’t enough to have a label ‘bullish’ or ‘bearish’ attached to a stock, but to have a discrete value belonging to an interval: for example of [-1;1] or [0;100]. Using such range of values, the user is able to quantify how bullish or bearish a sentiment is with regard to a particular stock, how volatile that sentiment is, how the fast and slow sentiment moving averages behave and so on. The range of values generation is another SSIX core features that is reflected within a set of social sentiment parameters (X-scores) generated based upon the SSIX raw sentiment data output.

Finally, back testing is required to validate the impact of using social sentiment for investment and trading. This means simulating the past using realistic conditions, ‘what-if’ investment and trading scenarios. Such exercise is achieved by building dedicated investment and trading strategies (decision and execution) to be run firstly with no sentiment X-scores involved (the benchmark set) and secondly with the sentiment X-scores added.

Figure 1 Backtesting example in MAARS platform, where a portfolio’s performance (green line) is compared with the S&P 500 index (red line). The lower graph shows the portfolio’s volatility and risk parameters evolution over the same period of time.

In the next article we’ll show how SSIX X-scores integrate within an investment strategy, how backtesting is executed and what influence social sentiment has in certain conditions over a portfolio’s performance.

This blog post was written by SSIX partner Laurentiu Vasiliu at Peracton Ltd.
For the latest update, like us on
Facebook, follow us on Twitter and join us on LinkedIn.