Projects
KIS Markets Financial Analysis - kismarkets.com
Problem: Statistically and visually analyzing recessions and market cycles is very difficult, tedious, and time consuming, especially
when comparing different recessions to one another in order to analyze recurring patterns and anomalies.
Action:
By developing and implementing algorithms based on Technical Analysis, we identify when recessions and cycles have occurred. We then use Intermarket Analysis for a comprehensive global view of each cycle's or recession's different phases. Next, we apply custom machine learning algorithms to examine recurring technical and intermarket patterns across these cycles. Finally, we present the results in a user-friendly and easily understandable manner.
Solution:
A platform that allows you to analyze both statistically and visually the differences and similarities between all recessions, cycles, and downturns
dating back to 1970 for all four major markets (Stocks, Bonds, FX, Commodities).
Fractal Asset Rotation
Problem: Currently, sector rotations are charted on a single timeframe, typically capturing only secondary trend rotations lasting 5-10 weeks. This limited view makes it challenging to discern whether a market move is a continuation, reversal, or correction relative to the longer-standing primary trend, which spans 1-3 years or more.
Solution:
The proposed solution is a visualization tool that simultaneously displays Primary (1-3 years), Secondary (4-12 weeks), and Daily rotations. This comprehensive view facilitates easy classification of whether a secondary rotation is aligning with or opposing the longer primary trend. It will also provide a big picture view of several assets all in one graph.
Correlation Data Extender
Problem:
When analyzing the historical performance of semiconductors leading up to every major recession or downturn, a challenge arises. Most semiconductor ETFs were created in the late 90s or early 2000s, but S&P 500 data is available from the early 1900s. This discrepancy means that semiconductor performance can only be compared for less than half of the entire dataset, rendering a significant portion of market data unusable.
Action:
To overcome this, identify all semiconductor-related stocks with data extending further back than the ETFs. Conduct a correlation test to determine which stocks have a correlation greater than 0.89. Choose the stock with the highest correlation (in this case, TXN with 0.975). Use the data from this stock (TXN), from its earliest data point to the inception of the semiconductor ETF (1992-2000), and apply a multiplier to each data point so that its final price aligns precisely with the ETF's initial price point. This adjusted dataset is then prepended to the original ETF dataset, creating a separate, extended asset.
Solution:
An algorithm that extends the data set for ETFs by retrieving all related stocks with longer data sets and selecting the stock with the highest correlaion over 0.89 to extend the ETFs data, thus providing a more comprehensive historical analysis.
Web Scrapper
Problem: Gathering historical asset data from reliable sources going back 50+ years is extremely expensive, hard to come by,
and tedious, particularly if you want the historical data of Stocks, Bonds, Currencies, Commodities, and Economic data.
Action:
First identify reputable financial websites with extensive archives. Use Python with libraries like Scrapy for parsing web pages.
Ensure your scraper navigates and extracts data like prices and dates efficiently. Respect site policies and avoid overloading servers.
Store the scraped data in a structured format, such as a database or Excel file, enabling easy retrieval and analysis. Regular updates and maintenance of the
scraper are essential.
Solution: A web scrapper that retrieves and formats historical asset data from numerous exchanges and reputable data sources
and stores it in a database that gets updated daily.
Data Science Supercomputer
Problem: Storing hundreds of terabytes of financial data and trying to run large complex financial models on a pc is impractical. One alternative is to use the cloud
to store data and run Machine Learning models and other simulations. However, they do not protect any intellectual property and can become very expensive over time.
Action: Source all of the neccessary parts to build a super fast Desktop computer and assemble.
Solution: A desktop computer that can run lightning fast machine learning models and other financial simulations while safely securing
intellectual property.
Specs: RTX 4080 + RTX 2080 GPUs | 128GB RAM | 10TB HDD | 2TB SSD | i9-13950HX (24 cores) CPU | Fedora OS
Crypto-Mining Racks
This project was a compelling fusion of my technical skills and financial insights, offering a hands-on experience with blockchain technology,
which is revolutionizing data security and digital transactions.
The journey began with meticulous research and procurement of the right components—high-performance GPUs,
a robust motherboard, efficient power supply units, and reliable cooling systems. This phase was critical in understanding
the hardware requirements for an optimized mining setup, deepening my knowledge in computer hardware and systems engineering.
I also developed a deeper understanding of blockchain technology and its implications for digital finance.
Analyzing the profitability of mining operations required me to leverage my background in financial analysis,
examining factors like electricity costs, hardware depreciation, and the volatile nature of cryptocurrency markets.
Secure Personal Laptop
Problem: Personal computers are highly susceptible to being hacked, putting your personal information at risk of being stolen and misused.
Solution: A fully encrypted personal computer equipped with a secure operating system designed to thwart network attacks and other forms of OS-specific attacks. Included is a travel router that functions as a personal router, encrypting all outgoing traffic and screening incoming network traffic for potential viruses and hacking attempts. Additionally, it offers a range of browsers with varying security levels, from completely anonymous and fully encrypted browsing to standard but relatively safe browsers like Chrome and Firefox, suitable for less sensitive online tasks.
Portable Digital Security
Problem: Sensitive information on your computer can be easily hacked or stolen without you even knowing. It's possible for bad actors to intercept
important and private communications via messages and emails between you and others.
Action: Download, install, and configure the Tails operating system on a USB key. Customize settings and security levels to match your needs.
Solution: An operating system that lives on a portable USB key and allows users to securely boot into any computer system. Provides anonymity and
enhanced encryption to ensure all documents and information is safe from being duplicated or stolen. Comes pre-installed with tools for encrypted emails and messaging, enhancing communication security.