Context Analytics (CA) is proud to serve as an exclusive partner of Stocktwits—a dynamic financial social media platform where investors share insights, opinions, and real-time commentary on a wide range of securities. Through this unique collaboration, CA has since expanded its coverage to include Indian securities (NSE).
Using our proprietary Natural Language Processing (NLP) technology, we analyze social sentiment and distill it into actionable factors that reflect both the tone and volume of conversations at the security level. One of our flagship products, the S-Factor feed, features the S-Score—a metric that quantifies the positivity or negativity of sentiment for each security.
To establish a historical baseline and test these metrics, Context Analytics pulled messages from Stocktwits and calculated S-Factors. The S-Score, one of many factors, is a Z-Score that detects sentiment from Stocktwits. An S-Score greater than 2 indicates that the conversation over the last 24 hours is 2 standard deviations more positive than the previous 20 days, suggesting a bullish outlook for the stock price. Conversely, an S-Score less than -2 reflects negative sentiment and a bearish outlook.
Additionally, users of our S-Factor feed can apply a volume filter on securities to ensure there is sufficient conversation feeding into the S-Score. For this blog, we are focusing on NSE securities with an S-Volume greater than or equal to 3—ensuring that at least three tweets contributed to the conversation in the previous 24 hours.
By applying the S-Score and S-Volume filter, we create simple long and short buckets, which are then combined into a theoretical long/short portfolio. Prior to market close each day, we allocate securities based on two different S-Volume thresholds: one version uses an S-Volume filter of 1 and the other uses a filter of 3. For both variants, securities with an S-Score above 2 are allocated to the long bucket, while those with an S-Score below –2 are allocated to the short bucket. The combined portfolios are rebalanced daily, with close-to-close returns calculated and compared against the Nifty50 baseline.
Our initial filter using an S-Volume threshold of ≥1 delivered promising results. Under this configuration, the strategy averaged 28 long positions and 5 short positions each day. While the long portfolio outperformed the NIFTY 50 by 25% cumulatively and 11% annually, the combined long/short portfolio under the S-Volume ≥1 filter was slightly less effective than the long positions alone.
Building on these promising returns, we then implemented a stricter filter with an S-Volume threshold of ≥3. With this filter, we allocated securities with an S-Score above 2 to the long bucket and those with an S-Score below –2 to the short bucket. Despite averaging only 10 long positions and 1–2 short positions daily, the long portfolio continued to outperform the NIFTY 50. Moreover, the short portfolio, which had been robustly negative (losing 67.5% since the beginning of 2023), helped enhance overall performance. As a result, the combined long/short portfolio achieved an impressive 295% gain over two years, while significantly improving upon the risk ratios of the NIFTY 50, all while trading on just 12 securities on average daily.
This blog demonstrates that sentiment data from Context Analytics and Stocktwits can be effectively leveraged for global trading strategies. Our approach has shown notable improvements over the NSE baseline of the NIFTY50 using only our Sentiment Score and Volume. For more information on our global coverage, please visit www.contextanalytics-ai.com.