09-04-2025
How AI Social Sentiment Analysis Is Changing Stock Price Predictions
Kirill Sagitov, founder of winner of "Young Entrepreneur of Russia, 2019" award, and author of books.
getty
Social networks have become one of the most significant sources of diverse information that is hard to quantify numerically. Millions of people discuss stock prices daily, despite not being directly involved in the stock market. Even those with limited knowledge of finance feel the impact of market fluctuations on the economy and their personal wealth. These emotions shape public opinion, which businesses must consider in their strategies.
Social sentiment analysis is used to study these processes. It assesses how people and society react to events related to securities. In the current reality, where a viral tweet can trigger a market crash, this has become a powerful and useful tool.
If Argentina's President Milley had utilized sentiment analysis, he might have avoided the impeachment threat after his reckless endorsement of the LIBRA cryptocurrency, which caused its price to plummet, as reported by In February 2025, small businesses that the cryptocurrency was intended to support lost billions in just hours.
Social analysis is a technology used to evaluate the emotional climate and reactions of internet users to specific events, often unrelated to the economy. Emotions influence people's thinking and actions, which is crucial for businesses. This information helps predict changes and can subtly influence them.
For example, after Donald Trump's re-election to a second term, his media strategy shifted. During his first term, as reports, he frequently praised the U.S. economy and the stock market's growth, but such posts have now virtually disappeared from his social media. This shift could point to his team possibly using sentiment analysis to guide decisions.
Previously, sentiment analysis was based on official reports and media publications. However, this approach has a significant drawback: The information is delayed. Social media, in contrast, provides instant coverage, but analysts struggle to process large data volumes in real time. AI helps solve this problem.
The primary sources of data for sentiment analysis include:
• Popular social networks.
• Finance-focused forums, like StockTwits and Yahoo Finance.
• Finance-related blogs.
• News platforms.
• Expert opinions and statements from public figures.
The last point is crucial because it generates news and sparks discussions among experts and the general public. Some leaders use social media to influence the stock prices of their companies. They carefully craft their posts to be interpreted in multiple ways, making it difficult to hold themselves accountable, even though attempts are made, as notes.
There are three key methods used in sentiment analysis:
• Natural language processing (NLP) can help extract information from text and messages.
• Sentiment categorization separates sentiments into 'positive,' 'negative' and 'neutral,' with scores assigned to each category.
• Identifying dominant emotions such as fear, anxiety, and greed within society is also important.
For example, reported that U.S. consumer sentiment dropped 10% in February compared to January due to growing pessimism among 62% of Americans. In December, they were optimistic about the new administration's ability to curb inflation, but this did not happen. A decline in consumer sentiment is a significant blow to the economy.
Common tools for sentiment analysis include:
• Google Cloud NLP for text analysis.
• IBM Watson for sentiment shifts through AI.
• FinSentS for extracting financial insights from sentiment data.
• for analyzing community sentiment.
Investors and traders need to react quickly to events that may prompt stockholders to buy or sell. For example, positive news about a company or reports of a catastrophic mistake by top management. In this case, acting fast is crucial, before people begin trading on the news.
AI that analyzes social sentiment helps by tracking changes and predicting how people will react in terms of buying or selling stock. This gives traders an edge, allowing them to forecast stock price movements more accurately.
Even AI-based analysis does not eliminate several challenges:
• Fraudulent Activity: estimates that 48% of all business-related messages contain false or irrelevant information.
• Discrepancies Between Sentiment And Actions: People may feel happy about something, but this doesn't guarantee they'll buy stock, especially if they lack the funds.
• Manipulation Of Sentiment: Disappointment after learning the truth can affect decisions.
• Human Factors: Recognizing sarcasm or humor can be difficult, especially across different languages and cultures.
• Short-Term Predictions: Sentiments are volatile, so sentiment analysis is unreliable for long-term investments.
As machine learning technologies improve, AI's prediction accuracy could also increase. AI will better understand human nature, emotions and motives.
The integration of alternative data sources, such as web traffic and app data, could help provide deeper insights into how people's sentiments are expressed, not just through social media posts but also through other activities.
The combination of sentiment analysis with blockchain technology could also help predict events on cryptocurrency exchanges, a rapidly growing field.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
