How to start Value Investing as a Data Scientist – DataDrivenInvestor | Candle Made Easy

The dos and don’ts of data science in value investing

Photo by David Pupaza on Unsplash

Granted, the only reason I got into data science was to make money.

I don’t have the Ph.D. or masters path because I seriously had no scientific curiosity about the field of data science.

Actually, I just wanted to be good with data to be a good investor.

So a degree in data science was really all I needed.

Also, to be honest, data science and value investing don’t really go hand-in-hand.

The reason for this is that machine learning works best in a closed and predictable environment; However, the stock market is unpredictable and driven by social, political and even environmental factors.

Perhaps you could gain an upper hand by using NLP on news articles or latent deep learning on sudden stock price changes, but leave that to the day traders.

But if you want to make value investing, should you ditch your data science skills and just start investing?

Definitely not!

Value investing is an art that takes years to perfect, and the feedback loop is long. However, your data science skills can make this long journey more productive.

1. Use your Python, R, or SQL skills to automate data wrangling for financial reports

Photo by JESHOOTS.COM on Unsplash

Value investing puts an emphasis on reading financial statements, and that’s a skill you need to learn separately.

However, you will soon run into the challenge that it is not easy to compare fiscal years as annual reports only ever provide data for 2 years in an annual report.

There are 3 ways to fix this.

To have several years of balance sheet data in one dataframe, you can call it from an API like Yahoo Finance and process the data to put it into a usable format. The only problem with this method is that free APIs are usually only limited to the last 4 years.

Another way I like to use it the most is to fill out a form while reading the annual report and selectively fill in certain financial values ​​on the form. You can then use Python to interpret the form’s data into something that can be interpreted.

The final method is to use computer vision to translate a financial report into a spreadsheet. I have found this to be an unreliable method, even when using home made and commercially made computer vision products. Problems arise when spaces in financial statements and accounting terminology change over time.

The takeaway point here is to experiment with different ways to automate the wrangling of financial reports into a useful format.

2. Use text-to-speech to listen to business reports

Most of the annual report is boring to read, except for the places where you find statistical anomalies and conduct an investigation.

However, it would be foolish of you to invest based solely on statistical analysis. It’s like taking care of a baby and believing that all it needs is nourishment. Feeding a baby allows them to grow healthily, but nurturing ensures a healthy, functioning baby.

That’s what reading the footnotes is for. The numbers can tell you if the company is healthy, but the footnotes can tell you if the managers are doing it in a sustainable way.

I’m pretty lazy so I use a text to speech app and copy and paste footnote text into the app.

If you’re cheap or have access to a good language training kit, you can run the text through a deep learning text-to-voice API to read the text to you.

For example, you could do this with the AWS Polly API.

Also, it makes no sense to apply NLP to annual reports. The authors of annual contributions, usually copy-and-paste explanations from previous years, and useful insights are mostly suppressed by statistically based NLP, since the fluff takes up significantly more space.

3. Perform data visualizations

Since you won’t be using large data sets, the next best way to identify patterns is through data visualization, also known as exploratory data analysis.

There are no fancy techniques here. To see if your company has become more profitable over the years, all you have to do is graph 10 years of EBIT data and find the slope of the best fitting line.

Here’s my rule of thumb.

The higher the level of the data set, the simpler the algorithm you need to use. For example, transactional data works with boosted trees, but financial summaries require a one-degree linear regression.

Also, I’d be lying if I told you to do this in Python. Since most value investing is one company at a time, it is more time efficient to plot some data in a spreadsheet and use the built-in graph builder to create the chart.

The only exception to this is after you’ve created multiple business analysis spreadsheets, you can use loops and functions to compare all business data at once.

Another way to do this is to make API calls to financial APIs and plot data in a chart. However, the weakness of this is that free APIs are usually limited in the number of calls you can make monthly and, as mentioned, you will most likely only have 4 years’ worth of data.

Conclusions: Does Data Science Make You a Value Investor?

In my honest opinion, yes it does.

What financial analysts lack is the ability to work more productively with the data at their disposal (e.g. finding better ways to make sense of the data) and what data scientists lack is an understanding of accounting and possibly a drive to Finding patterns where none exist.

However, with some persistence and effort, a data scientist can become a great value investor once they understand the data they are working with. I say that from my own experience. I used to be naive about investing, but now I’m cautious about investing and financials.

Leave a Comment