Investors managing multiple brokerage accounts often face challenges with inconsistent data formats, missing fundamental information, and a lack of interoperability between platforms. This is particularly relevant if you're tracking a portfolio across various brokers, aiming to centralize dividend income and position performance analysis, or building a personal finance dashboard without a dedicated data engineering team.
Historically, manual portfolio tracking involved a repetitive and time-consuming process. Each month, one would export CSVs from different brokers, encountering varying column names, date formats, and ticker conventions (e.g., AAPL vs. AAPL.US). The data cleaning phase required renaming columns, standardizing date formats, removing duplicate headers, reconciling tickers, and manually calculating cost basis and unrealized P&L. This often took two hours, rendering the data outdated by the time it was usable. Furthermore, broker exports rarely included crucial fundamental data like P/E ratios, dividend yields, EPS, or earnings dates, necessitating separate, ticker-by-ticker lookups.
The core issue was not the data itself, but the absence of an efficient system to connect and process it.
To address this pain point, the author moved away from manual spreadsheet manipulation and rebuilt the workflow around two key tools:
- Claude Cowork: Handles the messy human layer, encompassing files, formats, and logic.
- EODHD API: Manages the data layer, providing prices, fundamentals, and historical records.
Claude Cowork: From Chaos to Clean Data
Claude Cowork is Anthropic's desktop AI agent, capable of reading local files, understanding their structure, writing code, and executing it—all within a single session. Users simply drop broker export files into a folder. Cowork reads all three, identifies their respective schemas, normalizes column names, standardizes ticker formats, and merges everything into a unified dataset. It also categorizes each operation by type: buy, sell, dividend, or fee.
This eliminates manual cleaning or pre-scripting. Users describe their desired outcome, and Cowork determines the execution path, including writing and running the Python script for the actual data transformation. What once took two hours is now completed in under five minutes.
EODHD API: Enriching Every Position
Once the positions are cleaned and standardized by Cowork, the script calls the EODHD API to enrich each holding with data that brokerage exports never provide.
For every ticker in the portfolio, the script pulls:
- End-of-day price and historical performance
- Fundamental data: P/E ratio, EPS, dividend yield, payout ratio
- Company metadata: sector, market cap, exchange
- Upcoming dividend dates and amounts
The EODHD API covers over 70 exchanges and 150,000+ tickers, delivering consistent JSON responses that integrate cleanly into Python workflows. A single API key and endpoint structure provide all necessary data.
