Skip to main content

Features and Use Cases

Comprehensive analysis capabilities and monitoring tools for the ZettaQuant platform.

Core Analysis Features​

SEAL Analysis Framework​

Our proprietary SEAL (Sentence Evaluation and Analysis Logic) framework provides:

Financial Intelligence

  • Stance Analysis: Hawkish/dovish sentiment detection
  • Forward-looking Signals: Identify future-oriented statements
  • Uncertainty Measurement: Quantify ambiguity and confidence levels
  • Custom Metrics: Build domain-specific analysis variables

PDF Processing​

  • Automated text extraction and structuring
  • Batch processing of document archives
  • Format normalization and quality validation
  • Metadata enrichment and indexing

Technical Implementation​

Model Deployment​

  • Apply models across full datasets at scale
  • Efficient processing with GPU acceleration
  • Version control and model lineage tracking
  • A/B testing framework for model comparison

Data Integration​

  • Secure Storage: All results stored in your Snowflake database
  • Structured Output: Document-level metrics with sentence-level details
  • Versioned Results: Reproducible analysis with full audit trail
  • Custom Integration: Direct SQL access for downstream applications

Addendum​

Metrics & Analysis Pipeline

Our SEAL (Sentence Evaluation and Analysis Logic) pipeline processes your data through several sophisticated steps to generate actionable insights.

Pipeline Components​

1. Topic and Metric Definition​

  • Define relevant topics for filtering
  • Create custom metric variables
  • Set up classification parameters
  • Configure analysis framework

2. Efficient Sentence Labeling​

  • Cortex-powered weak labeling
  • Efficient sentence-level processing
  • Custom labeling models
  • Scalable analysis framework

3. Document-Level Metrics​

Generate comprehensive document-level metrics including:

  • Stance Analysis (hawkish/dovish)
  • Forward-looking Signals
  • Uncertainty Measurement
  • Custom Metric Construction

4. Downstream Analysis​

  • Perform quantitative analysis
  • Generate insights from labeled data
  • Integration with your analytics workflow
  • Ready for custom applications

Technical Implementation​

Model Application​

  • Apply models across full dataset
  • Efficient processing at scale
  • Secure computation within Snowflake

Data Storage​

  • Results stored in your Snowflake DB
  • Document-level metrics with labeled sentences
  • Versioned and reproducible outputs

Analysis Framework​

Our SCALAnalysis framework provides:

  • Customizable metrics
  • Flexible analysis options
  • Integration with your workflows
  • Reproducible results

Access Methods​

  • Streamlit UI for interactive analysis
  • Stored procedures for programmatic access
  • Direct SQL queries for custom analysis

title: Platform Overview​

Platform Overview

Our Snowflake Native App consists of two main components that work seamlessly within your Snowflake environment. All data and processing remain secure within your Snowflake infrastructure.

Key Features​

  • Native Snowflake Integration: All data resides in your Snowflake database
  • Real-Time Data Access: Near real-time data access through the native app
  • Custom Metrics: Define topics and metrics tailored to your needs
  • Secure Processing: All computations happen within your Snowflake environment

1. Data Selection & Ingestion​

The app supports three ways to work with data:

  1. Use Existing Data

    • Select document metadata tables
    • Choose tokenized sentences tables
    • Use data from your existing database and schema

  2. PDF Ingestion Tool

    • Drag-and-drop interface for PDFs
    • Configurable PDF parser
    • Custom tokenizer selection
    • Direct ingestion into Snowflake tables

  3. Premium Datasets

    • Real-time data updates
    • Premium access to complete datasets
    • Example:
      • Free version available on marketplace
      • Premium version with daily updates
      • Complete historical data

2. SEAL Pipeline Core Module​

Our core processing pipeline helps you:

  1. Topic Definition

    • Define relevant topics
    • Filter data based on relevancy
    • Custom topic modeling

  2. Classification & Metrics

    • Sentiment analysis
    • Stance detection
    • Uncertainty measurement
    • Custom metric construction

  3. Results Storage

    • Document-level metrics
    • Direct storage in your Snowflake DB
    • Ready for immediate analysis

Interface Options​

Users can interact with the app through:

  • Streamlit UI for visual interactions
  • Stored procedures for programmatic access

title: Data Ingestion​

Data Ingestion

Our native app provides multiple ways to ingest and work with data, all within your Snowflake environment.

Data Sources​

1. Customer's Own Data​

  • Upload unstructured data (PDF files, etc.)
  • Document ingestion through Native App component
  • Direct integration with your Snowflake DB
  • Secure compute within your environment

2. Premium Datasets​

  • Near real-time data access
  • Example:
    • Latest data available only through Native App
    • Daily updates
    • Complete historical coverage

3. Existing Snowflake Tables​

  • Use your pre-existing document metadata tables
  • Leverage already tokenized sentences
  • Direct integration with your schema

Ingestion Process​

  1. Document Processing

    • PDF parsing with configurable options
    • Document metadata extraction
    • Secure processing within Snowflake

  2. Text Tokenization

    • Configurable tokenizer selection
    • Sentence-level segmentation
    • Efficient storage in your Snowflake DB

  3. Data Storage

    • All processed data stays in your Snowflake environment
    • Optimized table structures
    • Ready for immediate analysis

Table Structures​

After ingestion, your data will be organized into two main tables:

Documents Table​

  • Document metadata
  • Source information
  • Timestamps
  • Document-level attributes

Sentences Table​

  • Tokenized sentences
  • Links to parent documents
  • Ready for analysis and labeling