About
This Swiss-based investment management firm specializing in alternative investing, with $95B AUM, sought to leverage natural language processing (NLP) to extract valuable insights from earnings call transcripts.
Analyzing textual data sourced from transcripts, articles, and documents can prove time-consuming and costly. With our NLP-ready datasets, you can skip the entire data management process and dive directly into analysis. Moreover, benefit from comprehensive metadata tagging and seamless integration with other datasets like Financials and Estimates to enrich your analysis and deliver critical insights.
Natural Language Processing (NLP) is a field of artificial intelligence (AI) that focuses on enabling computers to understand, interpret, and generate human language in a manner that is both meaningful and useful. It involves the development of algorithms and models that allow computers to process and analyze natural language data, such as text and speech, in order to extract meaning, identify patterns, and derive insights.
NLP tasks include parsing and analyzing text to extract information, identifying entities and relationships, translating text between languages, determining sentiment or emotion in text, converting speech to text, generating natural language responses, and summarizing text.
With our textual data suite, you can access machine-readable transcripts, filings, broker research and proprietary S&P Global credit research in structured formats for NLP analysis, so you can skip the entire process of sourcing, cleansing, and maintaining the data, and jump straight into insightful analysis. Read more in our Natural Language Processing Primer
Machine Readable Earnings Call Transcripts
The Machine Readable Transcripts dataset aggregates data from earnings calls delivered in a machine-readable format for Natural Language Processing (NLP) applications with metadata tagging.
Machine Readable Filings
Leverage Global Machine Readable Filings to perform Natural Language Processing (NLP) on an entity’s filings over time. Use the database to monitor strategic initiatives, justification of earnings, M&A plans, tactical execution, and much more.
Machine Readable Broker Reports
Machine Readable Broker Research unlocks the value within equity research reports by cleansing and parsing reports to deliver structured text from partner brokers for Natural Language Processing (NLP) applications
Textual Data Analytics: Sentiment Scores & Behavioral Metrics
The Textual Data Analytics (TDA) dataset takes earnings calls transcripts one step further with sentiment and behavioral-based metrics rigorously researched and tested against frequently used quantitative strategies.
This Swiss-based investment management firm specializing in alternative investing, with $95B AUM, sought to leverage natural language processing (NLP) to extract valuable insights from earnings call transcripts.
The team saw the benefits of using natural language processing (NLP) but gathering and maintaining the information as well as developing algorithms would require an extensive amount of time. In addition, the team was concerned about coverage, quality, and reliable data delivery on a daily basis.
Solution engineers from S&P Global recommended enhancing the current strategy mix with numerical scores obtained from earnings call transcripts. As the investment firm has been using the CIQ Financial data for building and managing its equity strategies, the new dataset was integrated smoothly into the existing investment framework. The scores are delivered in a structured format facilitating an efficient strategy backtest. The extensive meta-data allows for flexible score aggregation on the industry, geographical, and index levels. The scores are delivered within 90 minutes after the end of the earnings call which enables the investment team a timely reaction to any unexpected news.
The team gained access to a wide array of sentiment and behavioral scores for companies around the world, which were pre-tagged, structured, and organized. This allowed them the flexibility to combine the TDA scores with other investment data by leveraging the meta-level tags included in the package. Furthermore, they were able to operationally enhance pre-existing strategies with new textual investment signals quickly and efficiently. In addition, the speedy, intraday update of scores enabled them to respond promptly to unexpected news.
This research is part of a series from S&P Global's Quantamental Research team, which exclusively reviews earnings seasons using text from earnings call transcripts, providing insights generated from natural language processing algorithms.
READ THE REPORT
Our Quantamental Research group leverages the uniqueness and depth of our combined data and analytics offerings to uncover new investment insights. The award-winning research team is comprised of PhDs, CFAs, and former practitioners, and have helped clients around the globe make the most of their data.
DISCOVER MORE QUANTAMENTAL RESEARCH
Kensho NERD (Named Entity Recognition and Disambiguation) is a cutting-edge machine learning system that unlocks the full potential of your textual data by linking it to existing sources of structured knowledge. Trained on millions of business-related documents, NERD is the only technology on the market specifically optimized to extract financial entity information from text documents.
Kensho Scribe is a leading, on-demand transcription service for financial audio. With two complementary offerings available, Scribe AI and Scribe Human-in-the-Loop, your audio files are transcribed into human and machine readable text utilizing sophisticated deep learning models.
Scribe is purpose-built for financial audio, leveraging S&P Global's long history of providing high-quality transcripts to Wall Street, including more than 100,000 hours of domain-specific audio and associated text (e.g., earnings calls, management presentations, and acquisition announcements).
