How to Generate these Popular Stock Terms using Python

Sicong Zhao
6 min readJan 16, 2020

--

In the previous tutorial, we have introduced intrinsic value and 2 ways to download financial data (download here). Although very important, intrinsic value is not the only metric that value investors would consider.

In fact, a lot of important metrics are also easy to compute, which might serve as a better starting point especially if you do not feel confident about using Python or Pandas. But if you are already familiar with the tech and only care about intrinsic value, please feel free to skip to the next tutorial.

In this tutorial, we will cover the definition of some stock terms and then retrieve or calculate them from the financial data we have collected using Python. You can find the complete code at the end of this post.

The learning objectives are:

  • Learn what are some important terms for value investing
  • Learn to load financial data into Pandas
  • Learn how to retrieve and calculate the terms using Pandas

1. Stock Terms that will be covered

The following is the stock terms we will cover in this tutorial.

  • Shares Outstanding
  • EPS (Earning Per Share)
  • Dividend Rate
  • Dividend Yield
  • Debt/Equity Ratio
  • Book Value Per Share
  • ROE (Return on Equity)
  • Current Ratio
  • P/E (Price to Earning)
  • P/BV (Price to Book Value)

If you are already familiar with these terms, feel free to skip this part. If not, below is a video that covers all of these terms, you might want to watch it.

Just 2 quick notes: (1) The part about ‘Equity Growth’ (from 17:30 to 18:36) is not accurate. You might skip since it will not be covered in this section. (2) The definition of Current Ratio (from 20:27 to 22:23) is not correct, please refer to this Investopedia post.

2. Preparation

Codes are written in Python, you can use your local environment to do exercise. For the communication/demonstration purpose, I will upload my code to Google Colab and share it with you in this and next tutorial. The Google Colab is Jupyter Notebook running on Google’s server, so we do not need to worry about the compliance issue.

3. Load Data and Explore the Data

Pandas works like Excel. To do any calculation we need firstly load the data.

aapl = pd.read_csv('./stock_list/AAPL Key Ratios.csv', skiprows=2, index_col='Unnamed: 0')

In this line of code, ‘./stock_list/AAPL Key Ratios.csv’ is the file path, you need to substitute it with yours. Two parameters worth mentioning: ‘skiprows=2’ tells Pandas to ignore the first 2 rows. As you can see in fig.1 below, the first 2 rows are subtitles. Through ‘index_col=’Unnamed: 0'’, Pandas knows the first column should be index. (Yes, the column name ‘Unnamed: 0’ is annoying, but it just represents ‘the first unnamed column’. As you can see, if we ignore the first 2 rows, there is no column name specified for the first column.)

Fig.1 The structure of financial data

Then we can take a look at our dataframe:

aapl.head()

Using .head() function we can observe the first 5 rows of our data (Fig.2). It looks good.

Fig.2 Output of ‘appl.head()’

4. Get Terms From Data

It turns out the financial data we collected in the previous tutorial is very comprehensive, it contains most of the terms in our list. Let’s learn to retrieve them directly from our data, just so we can use them for further computation.

(1) Shares Outstanding

Let’s firstly get the Shares Outstanding in Sep. 2019. The corresponding index name is ‘Shares Mil’, which is an abbreviation of ‘Shares Outstanding in Million’. We can use function .loc() to locate the data. The first argument specifies the index, and the second specifies the column.

aapl.loc['Shares Mil', '2019-09']# output: 
# '4,649'

This result is fine except its format is string, not number. Which is not eligible for mathematical calculation. We can convert it to a number in the following 2 steps.

appl_share_str = aapl.loc['Shares Mil', '2019-09'].replace(',','')
appl_share = float(appl_share_str)
print(appl_share)
# Output:
# 4649.0

Instead of getting one data point, we might want the entire historical data. The following command would give us the complete historical ‘Shares Outstanding’.

aapl.loc['Shares Mil', ]# Output:
# 2010-09 6,473
# 2011-09 6,557
# 2012-09 6,617
# 2013-09 6,522
# 2014-09 6,123
# 2015-09 5,793
# 2016-09 5,500
# 2017-09 5,252
# 2018-09 5,000
# 2019-09 4,649
# TTM 4,649
# Name: Shares Mil, dtype: object

Compared with the code above, it simply leaves the second argument of loc() function blank.

In the output, ‘dtype’ means the data type. Similarly, we want it to be float for computational convenience. So, we can use the following code:

aapl.loc['Shares Mil', ].str.replace(',','').astype(float)# Output:
# 2010-09 6,473
# 2011-09 6,557
# 2012-09 6,617
# 2013-09 6,522
# 2014-09 6,123
# 2015-09 5,793
# 2016-09 5,500
# 2017-09 5,252
# 2018-09 5,000
# 2019-09 4,649
# TTM 4,649
# Name: Shares Mil, dtype: float64

(2) EPS (Earning Per Share)

We can do exactly the same thing as above. But in order to improve efficiency, let’s write a function and reuse it for the following terms.

def search_value(index_name, date):
return float(aapl.loc[index_name, date])
def historical_value(index_name):
return aapl.loc[index_name, ].astype(float)

So, we could use these two functions to retrieve data. The index name of EPS in our data is ‘Earnings Per Share USD’.

search_value('Earnings Per Share USD', '2019-09')# Output
# 11.89

The following code would update the aapl dataframe.

aapl.loc['Earnings Per Share USD'] = historical_value('Earnings Per Share USD')

(3) Other Terms

Below is a dictionary lists the index name of each stock term in our dataframe. I am not going to do the exercise for the other stock terms here, you are encouraged to do this exercise.

5. Calculate Stock Terms (P/E and P/BV)

Our data does not contain P/E (Price to Earning) and P/BV (Price to Book Value), so we calculate them.

(1) Get stock price

Based on the definition of these two terms, we need to get the stock price in order to compute them.

I have introduced IEX Cloud at the beginning, that’s a good choice. But here we use a package named yfinance to retrieve stock price because it’s free and easy to use.

Since yfinance is a 3rd party library, we need to install it before we use it. You can find documentation and the installation instruction here.

After installation, we can have our stock price in this way:

import yfinance as yf
apple = yf.Ticker('AAPL')
apple_price = apple.history(period='max')

We have load historical price data of Apple into apple_price in the form of Pandas dataframe. We can take a look using head() function.

apple_price.head()
Fig.3 The output of ‘apple_price.head()’

(2) Calculate P/E (Price to Earning)

Firstly, let’s consider what data to use for P/E. Earning is from the financial statement, which remains constant until the next financial report. Price changes every day. As we care mostly about the latest P/E, so we will calculate it using Earning Per Share from the latest financial records and yesterday’s close price for Apple.

latest_price = apple_price.loc['2020-01-14','Close']
latest_eps = search_value('Earnings Per Share USD', '2019-09')
latest_PtoE = latest_price/latest_eps
print(latest_PtoE)
# Output:
# 26.297729184188395

(3) Calculate P/BV (Price to Book Value)

Similarly, we care most about the latest P/BV, so we will calculate it using Book Value Per Share from the latest financial records and yesterday’s close price for Apple.

latest_eps = search_value('Book Value Per Share * USD', '2019-09')
latest_PtoBV = latest_price/latest_eps
print(latest_PtoBV)
# Output:
# 14.402579456471672

6. Summary

Thanks for reaching the end, I hope you have found something meaningful to you. If the Python or Pandas component in this tutorial annoyed you, maybe go back to the course description and use the materials listed to get more practice. If you feel easy about the content, then it’s a good time to start the next tutorial.

As mentioned above, I have pasted the code below. In the code, I also reformatted the data structure to make it more clean and readable. You can click the ‘Open in Colab’ button, and play around these codes online.

If you have any suggestions or feedbacks, please feel free to leave a comment or email me at sz163@duke.edu.

Previous tutorial: 1.Collecting financial data for fundamental analysis

Next tutorial: 3. How to calculate the intrinsic value

--

--

Sicong Zhao

Founder of Learniverse / Duke Alumni / Former Data Scientist @Credit Suisse / Former Computer Vision Fellow @Etsy /Former Product Designer @Baidu.