DeepAffects
Sentiment data on audio of earning calls transcripts, focusing on voice and emotion rather than content.
- Back-test and train models with data from 60,000 earnings calls.
- Recognize emotions on an 18 point scale, including frustration, excitement & more.around call topics such as revenue, sales, forecast, new-initiatives etc.
- Recognize emotions
Social/Sentiment
main data source2017
Year Company Founded- 1
Discretionary Asset Manager Customers10
Largest 3000 US public companies
$1MM
Hedge Fund
Agriculture, Airlines, Autos, Cannabis, China, Construction, Consumer, Energy, Financial Services, Gaming, Healthcare, Industrials, Insurance, Internet, Media, Real Estate, Restaurants & Food Delivery, Retail, Software (SaaS), Telecom, Travel
Milpitas, CA
Raw
Flat fee for historical data sets used for model training. Annual fee by # companies for real-time earnings call analysis.
2014
US
Far more accurate emotion analysis, which is direct from voice.
More accurate data, and more accurate data including identification of disagreements, optimism and caution. Calculate stress and emotions related to topics such as revenue, sales, forecast and new initiatives. Measure trends within topics across multiple earnings calls.
Asset trading companies interested to train their trading models, and increase their alpha with earnings call sentiments.