algotech-capital-management/algoseeknotebooks

Demo Notebooks for Quantitative Financee Using AlgoSeek

Algorithmic Trading with Python and AlgoSeek on AWS

Introduction

This is a collection of notebooks, recipes, and scripts demonstrating how to use AlgoSeek as a data provider for Quantitative Finance, Algorithmic Trading, and Machine Learning. There are samples covering everything from data ingestion (real-time and batch), stock/universe selection, backtesting, and feature engineering. Additionally, there are examples showing how to manipulate financial data with AWS EMR, PySpark, and AWS Sagemaker for end-to-end ML pipelines and intraday strategy research.

1) Overview

The notebooks in the root of this repo are the starting point. These are the broad introductions to sections detailed in the notebooks portion. The rest of the subdirectories are as follows

Directory	Description
algoseek	Library for algoseek-specific functions
data	data folder (gitignored but created in first notebook
Datasets	Samples and descriptions of AlgoSeeks Datasets
eda	Exploratory Data Analysis and Data Visualizations
integrations	Using AlgoSeek with external data sources
ML	Machine Learning Scripts
Notebooks	Miscellaneous Notebooks
Strategies	Trading Strategies
WIP	Work In Progress Notebooks

2) AlgoSeek Datasets

Here are dataset-specific notebooks exploring data for daily and intraday frequencies.

2.1) Equity Market Data:

These datasets contain the actual stock movement data.

Dataset	Description
BasicOHLCDaily
BasicAdjustedOHLCDaily
PrimaryOHLCDaily
PrimaryAdjustedOHLCDaily
StandardOHLCDaily
StandardAdjustedOHLCDaily
TradeAndQuote
TradeAndQuoteMinuteBar
TradeAndQuoteMinuteBarExcludingTRF
TradeOnly
TradeOnlyAdjusted
TradeOnlyAdjustedMinuteBar
TradeOnlyAdjustedMinuteBarBBG
TradeOnlyAdjustedMinuteBarExcludingTRF
TradeOnlyMinuteBar
TradeOnlyMinuteBarBBG
TradeOnlyMinuteBarExcludingTRF

2.2) Equity Reference Data:

These datasets provide more information about the securities and the Equity datasets.

Dataset	Description
BasicAdjustments
DetailedAdjustments
LookupBase
SecMasterBase

3) Data Access

There are currently two different ways to access the data: using the AlgoSeek SDK or using Boto3. There are notebooks for both methods here, but you should start with the introductions for both.

4) Time Series Analysis

5) Getting Started with Machine Learning

5.1) ML Preprocessing

5.2) Popular Libraries

Introductions to using AlgoSeek datasets with several popular libraries

5.3) ML Models

Some sample machine learning models to get you started

Model	Frequency	Description
Keras Univariate LSTM	Intraday	Intraday Regression
Linear Regression	Intraday
LightGBM Regression	Intraday
Random Forest Regressor	Intraday
XGBoost	Intraday

5.4) MLOps Framework

Model	Frequency	Description
LightGBM Regressor	Intraday	[Link](ml/mlflow/lightgbm/

5) Strategies

Strategy	Frequency	Description