-
Subscribe
Subscribed
Already have a WordPress.com account? Log in now.
> Launched
> In Development
> Tabled
> Aspirational
This website is a place to store my past projects so that I can reference and use them. Feel free to look around; all credit to third parties for inspiration or materials is given where necessary. If you feel that I should add a credit to your name on one of the pages, please feel free to reach out at the contact information below:
> Contact Info:
Projects on the site are chosen to match my interests and needs. The hope is that I will be able to use them either to learn new techniques or teach others. You will find that these projects are in various stages of completion and some might be returned to at a later date to add additional features.
# Launched
This is a paper I wrote as a sophomore for an REU program at IU Bloomington. As such, it is definitionally sophomoric. For many years, it has been hosted on their university website. However, I have fond memories of working with the biology and math departments to define the problem, learning probability and combinatorics over the summer, and working through the mathematics of compound probability distributions, pathways through the space of permutations of genes, and Monte Carlo simulations. I didn’t realize how valuable it was back then to be able to just sit and focus on a hard problem for an extended period of time.
Recent developments in LLM’s and protein folding make me feel like it might be worth revisiting this material at some point. The principal difficulties were the large number of genes involved and the size of permutation space. In any case, it is a fun trip down memory lane.
On occasion, I find myself teaching Python, usually while in the process of teaching differential equations, statistics, discrete mathematics, or business. This project uses a nice game that animates Python commands as it highlights the individual lines of code. My hope is both to create a resource that others can use to learn Python and also to demonstrate a little bit of my coding style with some visually pleasing examples.
In order to avoid spoiling the whole game, I won’t be presenting solutions to every challenge, and many other solutions (perhaps better ones) are possible. The main coding ideas illustrated are proper commenting, the development of functions as tools, the development of processes that use those tools, and the development of systems that use those processes.
Also, included will be examples of tuning parameters by minimizing the mean plus the standard deviation of runtimes.
> Tangent Portfolio Calculator:
This is an Excel sheet that takes in several stocks, computes the returns and variations for portfolios created from them, and assembles all of this information into a graph. This sheet was created in the process of helping Wake Forest students prepare for a presentation. I’m storing it here in case I need it for another student or class.
It can be useful to demonstrate how correlation between multiple stocks affects the optimal weight that we should assign to different stocks. For example, a stock might have a low return but be negatively correlated with a riskier stock. This lower return stock would then act to reduce the risk of the overall portfolio.
Swapping out stocks with different relationships allows us to see how the tangent portfolio will ‘leverage’ their relationships to achieve an optimal Sharpe ratio.
# Tabled
> Binomial Pricing of Options in Python
While researching this project, I came across a extremely well written post by Gregory Gundersen. I can’t imagine doing it better within any reasonable time frame. So, I’m going to table this project and link to his page so others can see his great work on the topic.
At some point, I may add some guardrails in the code or adapt them to demonstrate some Greeks to help students understand his material when I assign it. For now, however, please visit his page.
# In Development
> Ludum Dare:
Ludum Dare is Latin for “to give a game.” It has been a long running competition in which people create and submit a running video game using only 48 or 72 hours to complete their design, graphics, and programming. I’ve wanted to compete for years but haven’t had time due to other obligations. Perhaps ironically, the creator of this competition has recently had his own life events that have caused him to consider winding down this competition in October 2028. For that reason, this is a priority of mine as I always told myself that I would one day compete.
While I suspect that somebody else in the community will take up running a similar event, it would be nice to compete at least once officially.
# Aspirational
> LLM Lojban:
The growth in LLM’s, their harnesses, and the agents that utilize them has been very interesting to watch. However, their behavior is sometimes hard to understand, and the large number of model weights make it rater opaque. I would like to use a simplified language like Lojban to create my own Small Language Model (pun intended). The hope is to be able to see how tokens are processed and how the training data affects the weights on individual tokens.
https://mw.lojban.org/index.php?title=Lojban&setlang=en-US
This project is partly inspired by Ishan Anand’s successful and impressive attempt at running GPT-2 in a Microsoft Excel sheet.
GPT-2 gets squeezed into an Excel spreadsheet
Alternatively, I may use a small Markov chain to generate fake amino acid sequences with a start and stop codon. It would be easy to use such a system to develop gene “words” and a large number of them as well. It would be interesting to see if the SLM successfully approximates the Markov model used to generate these gene “words”.
> SOA Exam Excel Sheets:
I enjoy solving Actuarial Exam problems to maintain my skills. That said, now that the actual exams are over, I’m less interested in pushing some of the symbols around and finding rows in long tables and more interested in analyzing models and how different parameters affect different features of those models. These subject oriented Excel Sheets will create templates that can be used to set up and solve a wide variety of problems. That way, I can focus on correctly interpreting the problem as well as analyzing the results and the dependencies of those results on certain features. This will likely involve a liberal use of Excel’s Solver and XLOOKUP.
> DOS Business Calculator:
Microsoft has open sourced the original MS-DOS operating system for many years. I think it would be fun to try tackling programming from the opposite side of the spectrum from Python. As a small, hopefully manageable project, I hope to create a personalized business calculator and build it into a physical enclosure. The main objectives will be to improve the feel of the buttons by using mechanical keyboard switches, incorporate some lighting mechanics, and streamline my most common calculations.
Microsoft open sources the earliest dos source code discovered to date
> Finance Dashboard:
Both for my personal use and for the use of my finance students, I intend to use Power Query and Power Pivot along with perhaps some API calls to create a finance dashboard that will track the stock markets, the bond markets, LIBOR, Treasury, and mortgage rates, and a couple international markets. The Korean market is particularly interesting at the moment with its dominant memory companies and the leverage trading occurring on their stocks in 2026.
It would be nice to create a few macros that could execute stock strategies over historical data and calculate their performance and variation.
> Kaggle:
Having completed the Predictive Analytics and Statistical Risk Modeling SOA exams, I think it would be fun to begin participating in some Kaggle prediction competitions. They were mentioned in some of the reading material when highlighting the relative performance of Boosted Forests versus Random Forests.
> Excel Competition:
When I first heard about Excel competitions, it seemed delightfully absurd! I’ve watched a few of the competitors, and I don’t think I will ever reach their level. However, I think it would be fun to compete, and that I would learn a great deal in the process. I will likely start with some easier to enter competitions. If this sort of thing interests you, here are some links:
> Tangent Portfolio Calculator:
This is an Excel sheet that takes in several stocks, computes the returns and variations for portfolios created from them, and assembles all of this information into a graph. This sheet was created in the process of helping Wake Forest students prepare for a presentation. I’m storing it here in case I need it for another student or class.
It can be useful to demonstrate how correlation between multiple stocks affects the optimal weight that we should assign to different stocks. For example, a stock might have a low return but be negatively correlated with a riskier stock. This lower return stock would then act to reduce the risk of the overall portfolio.
Swapping out stocks with different relationships allows us to see how the tangent portfolio will ‘leverage’ their relationships to achieve an optimal Sharpe ratio.
Click to see more
> F1 Dashboard:
F1 commentary and data visualization has improved enormously over the years. The community also has a very organized and detailed set of data available on the Internet. I would like to take some time to create modern digital overlays for older races. Because the commentary style was less evolved some years ago, following these older races is actually quite difficult at times.
There are race commentators and race engineers that create and discuss their own graphical displays of data. They provide great insight into what goes on in their minds during the race when they are trying to make strategy calls. It would be interesting to recreate some of these charts and their lines of reasoning with older races.
You can find some f1 data generously collected and stored on this Github by toUpperCase78.
https://github.com/toUpperCase78/formula1-datasets
Ruth Buscombe, F1 commentator and motorsport engineer, has some wonderful data breakdowns of recent races that you can find on her channel.
Behance is a place where designers store their portfolio and recruiters can search for those designers. Some of these designers work directly with ggplot and data and are open with the code they used to produce that data. These designs are probably overkill for an actuarial report. However, they are beautiful and I think it would be a motivating way to learn more about TG plot and RS graphics.
Cédric Scherer Has some incredibly creative and beautiful data visualizations written in ggplot if that sort of thing interests you you can find his work in the link below
> 6502:
The 6502 was an incredibly important chip in the history of computers. It powered the Apple II, the Commodore PET, the Commodore 64, and the BBC Micro, among others. As a continuation of the theme of Excel Can Do Anything, I’d like to create my own logic gates in Excel and build out a working replica of a 6502 within a sheet.
Given the breath of this project, I’m going to start assembling the information that I will need below. you’ll find general information about 6502, directions to use 6502 to create your own computer, and circuit breakdowns of the internals of the chip below
Chip Hall of Fame: MOS Technology 6502 Microprocessor
https://github.com/davidmjc/6502
> Evergreen Projects 
