Sophisticated technologies such as Natural Language Processing (NLP) and Optical Character Recognition (OCR) have been increasingly used to create textual sentiment indicators in order to gauge the financial health of companies. These technologies are especially helpful in markets such as China where there is a lack of reliable and timely financial data.
See how S&P Global Data Scientist, Andrew Bovey, used the China Sentiment Analytics dataset with the Marketplace Workbench to create textual sentiment indicators for stock price prediction in the Chinese market. In this video, we will preview:
- How to extract the raw text (in simplified Chinese) from financial disclosures section and break into sub-components for analysis
- Calculating sentiment scores at a section, article, and company level
- Potential causal relationships between Chinese stock price movements and their underlying sentiment data
- Whether or not the China Sentiment Analytics dataset can help us predict future values of a separate signal
