How to Begin Value Investing as a Data Scientist
Written on
Chapter 1: Introduction to Value Investing
To be honest, my entry into data science was primarily driven by the desire to generate income. I didn't pursue a Ph.D. or Master's degree because I lacked a genuine scientific curiosity about the field. My main goal was to acquire data skills that would enhance my investment strategies. Therefore, a data science certificate was sufficient for my needs.
However, it’s important to note that data science and value investing don't naturally align. Machine learning thrives in structured and predictable environments, while the stock market is inherently unpredictable, influenced by a variety of social, political, and environmental factors. While techniques like natural language processing (NLP) applied to news or deep learning for sudden price changes may be useful for day trading, they aren't as applicable for long-term value investing.
So, should data scientists abandon their skills in favor of pure investing? Absolutely not! Mastering value investing is an art that requires years of practice, and your data science expertise can significantly enhance this journey.
Section 1.1: Automating Financial Data Analysis
Value investing relies heavily on analyzing financial statements, a skill that requires separate training. One challenge is that annual reports typically provide data for only two years, making comparisons difficult. Here are three approaches to tackle this issue:
- Utilizing APIs: You can gather multiple years of financial data into a single data frame by calling it from APIs like Yahoo Finance. However, be aware that free APIs usually restrict access to just the latest four years.
- Manual Data Entry: My preferred method involves filling out a form as I read through the annual report, selectively recording key financial figures. Then, I use Python to manipulate this data into a usable format.
- Computer Vision: Another option is to use computer vision technology to convert financial statements into tables. However, this method can be unreliable due to issues such as varying accounting terminology and formatting inconsistencies.
Ultimately, it’s beneficial to experiment with different methods for automating financial data wrangling.
Section 1.2: Enhancing Report Analysis with Technology
Most annual reports can be tedious to read, aside from sections that reveal statistical anomalies worth investigating. Relying solely on statistical analysis is akin to caring for a baby with just food; while nourishment is essential, nurturing is equally important.
Footnotes in reports provide critical insights into whether a company is being managed sustainably. To avoid the boredom of reading, I use a text-to-speech app to listen to the footnotes. Alternatively, if you're tech-savvy, you could leverage a deep learning text-to-speech API, like AWS Polly, to read the text aloud.
It's worth noting that applying NLP to annual reports may not yield meaningful insights, as writers often recycle past explanations, diluting the value of new information.
Chapter 2: Visualizing Financial Data
Since large datasets aren’t typically used in value investing, data visualization becomes a key tool for identifying patterns. A straightforward technique is to plot ten years of EBIT data to analyze profitability trends.
My guideline is this: the broader the data set, the simpler the algorithm required. For instance, while transactional data may benefit from complex algorithms, financial summaries often only need linear regression. I recommend using spreadsheets for individual company analyses, as the built-in graphing tools offer a quick and effective way to visualize data.
If you're analyzing multiple companies, once you've compiled several spreadsheets, you can utilize loops and functions in Python to compare data efficiently.
Concluding Thoughts: The Intersection of Data Science and Value Investing
In my view, data science can indeed enhance one's ability to be a successful value investor. Financial analysts often struggle to maximize their data productivity, while data scientists may lack a solid understanding of accounting principles. With dedication and effort, a data scientist can evolve into a proficient value investor, as I have experienced firsthand. Previously, I approached investing naively, but I've since learned to be more discerning when it comes to financial data.
If you found this article helpful, consider joining Medium for access to a wealth of articles from talented writers. Signing up through my link supports my work.