Portfolio Projects

Exploring patterns and insights through quantitative research and trading.

Geopolitics Sentiment Google Indicator

This project develops a Geopolitics Google Sentiment Indicator, addressing the growing impact of geopolitics on investment decisions amidst rising dollar index and global policy uncertainties. Using Google Trends data for key geopolitical terms, the algorithm scrapes, cleans, and processes search volume values (SVI), applying winsorization, deseasonalization, and normalisation. It incorporates paid proxy servers to manage complex data extraction and demonstrates a positive correlation with the Geopolitical Risk Index from Caldera and Iacoviello's research. The algorithm supports both weekly and monthly analysis, with complete data extraction, cleaning, and modelling workflows included. The algorithm is customisable as per user preferences.

NZ mutual funds ESG analysis

This project provides an in-depth analysis of New Zealand mutual funds, focusing on ESG and non-ESG portfolios. It includes data cleaning, organisation, and integration of fund-level and holdings-level datasets, covering metrics like returns, ESG scores, fund flows, and performance benchmarks. Key calculations include ESG tilts, flows, risk measures, alphas, volatility, and comparative studies between ESG and non-ESG portfolios. The analysis offers comprehensive insights into fund performance, trading behaviours, and ESG-driven investment trends.

More projects to be uploaded soon...

Yahoo Finance data scraping project

Yahoo Finance serves as a valuable and reliable platform for accessing the latest financial data at no cost. In this project, I leveraged the capabilities of yfinance to download a diverse array of datasets, encompassing stock and index price and volume data, along with fundamental financial statements such as balance sheets, cash flow and income statements from previous years and recent quarters across various markets. This comprehensive dataset not only provides insights into market trends but also serves as a critical resource for multiple financial analysis and trading projects. By utilising this data, investors and analysts can make informed decisions and strategies in an ever-evolving financial landscape.

Building financial data pipeline with SEC EDGAR filings

I developed a keyword identification bot specifically designed for SEC Edgar 10-K and 10-Q filings to enhance data extraction for my current job. In addition, I created a master index file that consolidates company filings from 1996 to the present and compiled insider trading data from 2006 to today for historical analysis. Given the vast size of the Edgar database, this project is still a work in progress. The repository is private, please contact me if you would like access to the consolidated data and the code I've developed. Furthermore, I implemented a data pipeline to facilitate text keyword search algorithms within the dataset, streamlining the process of extracting valuable insights from the extensive filings.