Lecture Notes: AWS Trading Bot Data Pipeline Tutorial

Introduction

• Overview of data pipeline for trading bot using AWS Cloud.
• Tutorial will cover interaction with Yahoo Finance API and DynamoDB.
• Steps include creating AWS user, setting up DynamoDB, configuring AWS client, running Python code, and testing.

AWS User Creation

• AWS User Creation Steps
  • Access AWS console, go to IAM service.
  • Create a new user with necessary permissions, mainly DynamoDB full access.
  • Generate access and secret keys.

DynamoDB Setup

• Table Creation
  • Create a new table, "Price History Table," with partition key (ticker) and sort key (timestamp).
  • Consider enabling deletion protection.

AWS Client Configuration

• Package Installation
  • Ensure boto3 and related packages are installed.
• Configuration
  • Configure AWS client with access key, secret key, default region, and JSON output format.

Python Code Overview

• Warm-Up Asset Data Function
  • Populate DynamoDB with maximum data from Yahoo Finance API if initially empty.
• Update Price Table Function
  • Update existing data in DynamoDB by comparing latest record date with current date.
• Database Operations Class
  • Methods for querying last records, converting data types, and inserting data.
• Requirements Clarification
  • List necessary operations like querying API, formatting data, and updating tables.

Code Implementation

• Imports and Classes
  • Key imports: boto3, yFinance.
  • Define classes to handle API interaction and database operations.
• Main Functions
  • Warm-up asset data and update price table functions handle data fetching and updating.

Testing and Debugging

• Testing
  • Unit tests available on GitHub to ensure code stability and robustness.
• Error Handling
  • Manage errors in database operations with conditional expressions.

Strategy Implementation (Part 2)

• Market Neutral Strategy
  • Pairs trading strategy using ETFs and inverse ETFs for market neutrality.
• Strategy Performance Analysis
  • Analysis of performance based on market conditions such as volatility and direction.
• Backtesting and Improvements
  • Backtesting results show correlation with S&P 500, potential improvements involve better portfolio weighting schemes.

Conclusion

• Overall framework is modular, allowing substitution of different APIs or databases.
• Strategy implementation details and backtesting available on GitHub.
• Suggestions for future improvements and further research avenues.