Topic Deltas

Discover hidden nuances in earnings call topics that move stocks

Contact Us
Image Description
Key Benefits

Topic Deltas reads through earnings transcripts, corporate filings and research notes in a granular bottom-up way measuring topical sentiment changes across calls and between industry peers, giving you the unique knowledge that scales your expertise.

Book a demo


Hours of tedious analysis into minutes with super-human accuracy



Key themes by comparing and contrasting granular topics over time and against those referenced by industry peers



The semantic details of textured, nuanced language to find patterns that were previously invisible



Predictive insights into company fundamentals by allowing Topic Deltas to read, understand and learn in specialized ways

Why Topic Deltas?

Supercharge your analytical workflow from the bottom-up with super-human comprehension

Instantaneously visualize topical sentiment on a company-specific basis and be the first to discover macro trends affecting entire industries.

  • Compare companies within an industry by topic

  • Construct macro level views of topical trends by industry

Topic Deltas makes it easier than ever to contrast statements made during earnings calls on a per-topic basis, against market expectations, over time, and at scale.

  • Human level semantic accuracy in measuring sentiment scoring and classification

  • Data extraction, transformation, categorization, and attribution

How it works

To accurately measure sentiment, a machine has to understand. To understand, the machine has to learn.

Modular Core Capabilities

Topic Deltas is built from a specific combination of our core technologies trained to understand the context of industry-specific language. It combines proprietary algorithms for metadata and snippet extraction, labeling, content enrichment, machine learning and deep learning, activation functions, and sentiment analysis to create dynamic continuously learning models that produce differentiated outputs yielding predictive value.

Image Description

Industry Specific Training

Industry analysts who understand key topics and related nuances create semantically rich training datasets on the order of less than .1 percent of the data from which Topic Deltas learns, thus scaling expertise. By understanding industry context in an adaptive manner, Topic Deltas is able to accurately compute sentiment with a high level of semantic accuracy. It learns and understands granular concepts like "crack spreads" and "revPAR" and a continuously expanding topical coverage and depth of understanding.

SVG Illustration

Continuous Learning and Improvement

The core technology powering Topic Deltas has the unique ability to adapt dynamically and incorporate new information and contexts as they arise. Since our systems continuously learn, improve, and builds of new insights, we expect the value of Topic Deltas to increase over time as it uncovers deeper insights from higher value data sources and develops a richer understanding across a variety of domains.

SVG Illustration

Related Content


Research Paper

Learn how JPMorgan Chase's research team used Topic Deltas sentiment to predict fundamental outcomes in European banks

Read More


What A.I. can and can't do for your investment portfolio

Read More

Uncover hidden nuance

See how earning calls sentiment shifts have moved stocks

Book a demo