Home
Hi, I'm Håkon M. Holmen.
I'm a self taught Data Scientist and Software Developer who is looking to turn my hobby into a career.
I started learning programming in 2019 after struggling to do some (very simple) financial predictive modeling in Excel. This quickly snowballed. And while this first small modeling project of mine was a failure (of course), I've since expanded my toolbox to include a wide range of tools and techniques.
I usually work with Python, but I'm also familiar with R, SQL, and Cython. I have a strong knowledge of the standard data science and machine learning stack (Numpy, Pandas, Scikit-learn, Torch, etc), and most of the common data visualization tools (Matplotlib, Bokeh, Plotly, etc). I'm also well versed in software development tools and best practices (version control, testing, design patterns, code documentation, etc).
Below you can find a list of my technical skills, as well as an overview of my technical interests and some of my recent projects.
Interests
As mentioned above, I started learning programming and data science due to an interest in finance and the economy. What captivates me about finance is the combination of the problem complexity and the real-world feedback from the market. You need a combination of erudition, rigorous thinking, and practical wisdom to succeed (neither of which I have). And the market is a great way to get real-world, non-BS feedback on your big idea. It will happily humble you when your big idea starts to delude you.
I would say that most of my other technical interests flew from my initial interest in finance. They are different tools for helping me understand how things work. Although some interests, like building something in Python, have grown into ends in and of themselves.
That said, here's an overview of the areas I find most interesting:
Finance
The problem complexity combined with the real world, non-BS feedback of the market is what makes finance so captivating.
Python
I sometimes find building something in Python, a tool, a chess-AI, a simple script, or really anything, a joy unto itself.
Probability Theory
Although I'm (still) not exactly an master of the craft, I find the field of probability theory the most important and useful tool for understanding the world.
Statistical Modeling
The tools and techniques for modeling real world phenomena. Extremely useful if applied correctly, which is often more difficult than the average practitioner likes to admit.
NLP
The ability to convert human language into legible information is both useful and interesting. My favorite use-cases are measuring the evolution of thoughts over time and measuring differences in thinking between groups.
Data Visualization
Graphs, plots, and dashboards are a great way to explain and summarize information. Simplicity is key to understanding, and a well-crafted plot is a key simplifier of data.
Skills
If I had to assess, I would consider the technical skills I've learned over the last couple of years to be on par with those of a starting junior data scientist or software engineer. However, this is difficult to ascertain as I've never had the chance to work alongside other data scientists or software engineers.
I possess a firm grasp of the key tools and libraries in data science, and have a decent understanding of the fundamentals of the field like statistics, data engineering, machine learning algorithms, etc. I am also proficient in software development best practices, such as version control, testing, design patterns, code documentation, clean code principles, etc.
I would consider my main weakness to be a lack of experience with big data technologies and enterprise software development tools such as Spark, Hadoop, Azure, Kubernetes, etc. These are tools I've still not had the chance to work with, mostly due to the fact that what I've learned originated from personal exploration and hobbyist pursuits.
Here's a simple overview of my current technical skills:
Languages
Python - 5
R - 2
Cython - 2
Libraries
DS (Pandas / Numpy / Polars) - 5
Viz (Matplotlib / Bokeh / Plotly) - 4
ML (Sklearn / XGBoost) - 4
NLP (Torch / Spacy / 🤗) - 4
Scraping (Scrapy / Tornado) - 3
Other
Version Control (Git) - 5
Databases (SQL) - 4
Linux / Bash - 3
Probability & Statistics - 3
My Projects
Here's an overview of some of my recent (published) projects:
New York Times Sentiment Index
An economic sentiment index based on New York Times headlines from 1896 till today.
DistilNews
Two robust and highly accurate transformer models for news classification trained a large, ChatGPT labeled dataset.
FeatherStore
Blazingly fast datastore for dataframes built upon Arrow Feather.
Contact
Feel free to contact me for any question. For open source projects, please open an issue or pull request on Github. Otherwise, you can reach out through any platforms below:
