Trading & Market Data Hub

Overview

Trading desks receive market data from multiple dealers in inconsistent formats — Excel spreadsheets, PDFs, emails, and CSV files with different column layouts, spread conventions, and rating scales. This platform centralizes ingestion, normalizes the data, and presents a unified view that eliminates hours of manual data preparation.

Technical Approach

graph TD
    A[Dealer Feeds] --> B[Format Detection]
    B --> C1[Excel Parser]
    B --> C2[PDF Parser]
    B --> C3[Email Parser]
    B --> C4[CSV Parser]
    C1 --> D[Normalization<br/>Engine]
    C2 --> D
    C3 --> D
    C4 --> D
    D --> E[Spread Convention<br/>Harmonization]
    E --> F[Rating<br/>Harmonization]
    F --> G[Time-Series<br/>Aggregation]
    G --> H[React<br/>Dashboard]

Multi-Format Ingestion: The platform automatically detects incoming file formats and routes them to specialized parsers. Each parser handles format-specific challenges — merged cells in Excel, table extraction from PDFs, attachment handling from emails.

Normalization Engine: Dealer-specific column mappings, spread conventions (OAS, DM, Z-spread), and rating scales (Moody’s, S&P, Fitch) are harmonized to a common schema. Configurable mapping files allow quick onboarding of new dealers.

Time-Series Aggregation: Historical data points are aligned to a common timeline, enabling cross-dealer comparison and trend analysis. Analysts can track how dealer pricing evolves over time for specific securities.

Key Results

• Ingests data from 8+ dealer feeds across 4 file formats
• Reduces analyst data preparation time by approximately 50%
• Harmonizes 3 rating scales and multiple spread conventions automatically
• Historical time-series enables cross-dealer pricing comparison

Technologies Used

Python React PySpark openpyxl Material-UI