Added paper 01C.

falcaoth · Dec 22, 2018 · faa29e7 · faa29e7
1 parent 7b9c584
commit faa29e7
Show file tree

Hide file tree

Showing 16 changed files with 3,794 additions and 46 deletions.
diff --git a/01C_Theory_of_Long-Term_Stock_Forecasting.ipynb b/01C_Theory_of_Long-Term_Stock_Forecasting.ipynb
diff --git a/README.md b/README.md
@@ -18,6 +18,8 @@ modified and run again.
 
 1-B. Better Long-Term Stock Forecasts ([Notebook](https://github.com/Hvass-Labs/FinanceOps/blob/master/01B_Better_Long-Term_Stock_Forecasts.ipynb)) ([Google Colab](https://colab.research.google.com/github/Hvass-Labs/FinanceOps/blob/master/01B_Better_Long-Term_Stock_Forecasts.ipynb))
 
+1-C. Theory of Long-Term Stock Forecasting ([Notebook](https://github.com/Hvass-Labs/FinanceOps/blob/master/01C_Theory_of_Long-Term_Stock_Forecasting.ipynb)) ([Google Colab](https://colab.research.google.com/github/Hvass-Labs/FinanceOps/blob/master/01C_Theory_of_Long-Term_Stock_Forecasting.ipynb))
+
 2. Comparing Stock Indices ([Notebook](https://github.com/Hvass-Labs/FinanceOps/blob/master/02_Comparing_Stock_Indices.ipynb)) ([Google Colab](https://colab.research.google.com/github/Hvass-Labs/FinanceOps/blob/master/02_Comparing_Stock_Indices.ipynb))
 
 3. Portfolio Optimization Using Signals ([Notebook](https://github.com/Hvass-Labs/FinanceOps/blob/master/03_Portfolio_Optimization_Using_Signals.ipynb)) ([Google Colab](https://colab.research.google.com/github/Hvass-Labs/FinanceOps/blob/master/03_Portfolio_Optimization_Using_Signals.ipynb))

diff --git a/data.py b/data.py
@@ -112,6 +112,122 @@ def _load_price_yahoo(ticker):
     return price_daily
 
 
+def _load_earnings_per_share(ticker, df, profit_margin=True):
+    """
+    Load the Earnings Per Share from a data-file and add it to the DataFrame.
+    Also calculate the P/E ratio and profit margin.
+
+    :param ticker:
+        Name of the stock used in the filenames e.g. "WMT"
+
+    :param df:
+        Pandas DataFrame with SHARE_PRICE.
+
+    :param profit_margin:
+        Boolean whether to add the profit margin to the DataFrame.
+        Requires that df already contains SALES_PER_SHARE.
+
+    :return:
+        None. Data is added to the `df` DataFrame.
+    """
+
+    # Load data.
+    path = os.path.join(data_dir, ticker + " Earnings Per Share.txt")
+    earnings_per_share = _load_data(path=path)
+
+    # Add to the DataFrame (interpolated daily).
+    df[EARNINGS_PER_SHARE] = _resample_daily(earnings_per_share)
+
+    # Add valuation ratio to the DataFrame (daily).
+    df[PE] = df[SHARE_PRICE] / df[EARNINGS_PER_SHARE]
+
+    # Add profit margin to the DataFrame (daily).
+    if profit_margin:
+        df[PROFIT_MARGIN] = df[EARNINGS_PER_SHARE] / df[SALES_PER_SHARE]
+
+
+def _load_sales_per_share(ticker, df):
+    """
+    Load the Sales Per Share from a data-file and add it to the DataFrame.
+    Also calculate the P/Sales ratio and one-year growth in Sales Per Share.
+
+    :param ticker:
+        Name of the stock used in the filenames e.g. "WMT"
+
+    :param df:
+        Pandas DataFrame with SHARE_PRICE.
+
+    :return:
+        None. Data is added to the `df` DataFrame.
+    """
+
+    # Load data.
+    path = os.path.join(data_dir, ticker + " Sales Per Share.txt")
+    sales_per_share = _load_data(path=path)
+
+    # Add to the DataFrame (interpolated daily).
+    df[SALES_PER_SHARE] = _resample_daily(sales_per_share)
+
+    # Add valuation ratio to the DataFrame (daily).
+    df[PSALES] = df[SHARE_PRICE] / df[SALES_PER_SHARE]
+
+    # Add growth to the DataFrame (daily).
+    df[SALES_GROWTH] = df[SALES_PER_SHARE].pct_change(periods=365)
+
+
+def _load_book_value_per_share(ticker, df):
+    """
+    Load the Book-Value Per Share from a data-file and add it to the DataFrame.
+    Also calculate the P/Book ratio.
+
+    :param ticker:
+        Name of the stock used in the filenames e.g. "WMT"
+
+    :param df:
+        Pandas DataFrame with SHARE_PRICE.
+
+    :return:
+        None. Data is added to the `df` DataFrame.
+    """
+
+    # Load data.
+    path = os.path.join(data_dir, ticker + " Book-Value Per Share.txt")
+    book_value_per_share = _load_data(path=path)
+
+    # Add to the DataFrame (interpolated daily).
+    df[BOOK_VALUE_PER_SHARE] = _resample_daily(book_value_per_share)
+
+    # Add valuation ratio to the DataFrame (daily).
+    df[PBOOK] = df[SHARE_PRICE] / df[BOOK_VALUE_PER_SHARE]
+
+
+def _load_dividend_TTM(ticker, df):
+    """
+    Load the Dividend Per Share TTM (Trailing Twelve Months) from a data-file and
+    add it to the DataFrame. Also calculate some related valuation ratios.
+
+    :param ticker:
+        Name of the stock-index used in the filenames e.g. "S&P 500"
+
+    :param df:
+        Pandas DataFrame with SHARE_PRICE.
+
+    :return:
+        None. Data is added to the `df` DataFrame.
+    """
+
+    # Load data.
+    path = os.path.join(data_dir, ticker + " Dividend Per Share TTM.txt")
+    dividend_per_share_TTM = _load_data(path=path)
+
+    # Add to the DataFrame (interpolated daily).
+    df[DIVIDEND_TTM] = _resample_daily(dividend_per_share_TTM)
+
+    # Add valuation ratios to the DataFrame (daily).
+    df[PDIVIDEND] = df[SHARE_PRICE] / df[DIVIDEND_TTM]
+    df[DIVIDEND_YIELD] = df[DIVIDEND_TTM] / df[SHARE_PRICE]
+
+
 ########################################################################
 # Public functions.
 
@@ -214,7 +330,7 @@ def load_index_data(ticker, sales=True, book_value=True, dividend_TTM=True):
     path = os.path.join(data_dir, ticker + " Dividend Per Share.txt")
     dividend_per_share = _load_data(path=path)
 
-    # Merge price and dividend into a single data-frame.
+    # Merge price and dividend into a single DataFrame.
     df = pd.concat([price_daily, dividend_per_share], axis=1)
 
     # Only keep the rows where the share-price is defined.
@@ -225,82 +341,68 @@ def load_index_data(ticker, sales=True, book_value=True, dividend_TTM=True):
     # for stock indices because it does not reinvest dividends.
     df[TOTAL_RETURN] = total_return(df=df)
 
+    # Load Sales Per Share data.
     if sales:
-        # Load Sales Per Share data.
-        path = os.path.join(data_dir, ticker + " Sales Per Share.txt")
-        sales_per_share = _load_data(path=path)
-
-        # Add to the data-frame (interpolated daily).
-        df[SALES_PER_SHARE] = _resample_daily(sales_per_share)
-
-        # Add P/Sales ratio to the data-frame (daily).
-        df[PSALES] = df[SHARE_PRICE] / df[SALES_PER_SHARE]
+        _load_sales_per_share(ticker=ticker, df=df)
 
+    # Load Book-Value Per Share data.
     if book_value:
-        # Load Book-Value Per Share data.
-        path = os.path.join(data_dir, ticker + " Book-Value Per Share.txt")
-        book_value_per_share = _load_data(path=path)
-
-        # Add to the data-frame (interpolated daily).
-        df[BOOK_VALUE_PER_SHARE] = _resample_daily(book_value_per_share)
-
-        # Add P/Book to the data-frame (daily).
-        df[PBOOK] = df[SHARE_PRICE] / df[BOOK_VALUE_PER_SHARE]
+        _load_book_value_per_share(ticker=ticker, df=df)
 
+    # Load Dividend Per Share TTM data.
     if dividend_TTM:
-        # Load Dividend Per Share TTM data.
-        path = os.path.join(data_dir, ticker + " Dividend Per Share TTM.txt")
-        dividend_per_share_TTM = _load_data(path=path)
-
-        # Add to the data-frame (interpolated daily).
-        df[DIVIDEND_TTM] = _resample_daily(dividend_per_share_TTM)
-
-        # Add Dividend Yield to the data-frame (daily).
-        df[DIVIDEND_YIELD] = df[DIVIDEND_TTM] / df[SHARE_PRICE]
-
-        # Add P/Dividend to the data-frame (daily).
-        df[PDIVIDEND] = df[SHARE_PRICE] / df[DIVIDEND_TTM]
+        _load_dividend_TTM(ticker=ticker, df=df)
 
     return df
 
 
-def load_stock_data(ticker):
+def load_stock_data(ticker, earnings=True, sales=True, book_value=True):
     """
     Load data for a single stock from several different files
     and combine them into a single Pandas DataFrame.
 
     - Price is loaded from a Yahoo-file.
-    - Sales Per Share and Book-Value Per Share are loaded from separate files.
+    - Other data is loaded from separate files.
 
     The Total Return is taken directly from the Yahoo price-data.
-    The P/Sales and P/Book ratios are calculated daily.
+    Valuation ratios such as P/E and P/Sales are calculated daily
+    from interpolated data.
 
     :param ticker:
         Name of the stock used in the filenames e.g. "WMT"
+
+    :param earnings:
+        Boolean whether to load data-file for Earnings Per Share.
+
+    :param sales:
+        Boolean whether to load data-file for Sales Per Share.
+
+    :param book_value:
+        Boolean whether to load data-file for Book-Value Per Share.
+
     :return: Pandas DataFrame with the data.
     """
-    # Paths for the data-files.
-    path_sales_per_share = os.path.join(data_dir, ticker + " Sales Per Share.txt")
-    path_book_value_per_share = os.path.join(data_dir, ticker + " Book-Value Per Share.txt")
-
     # Load the data-files.
     price_daily = _load_price_yahoo(ticker=ticker)
-    sales_per_share = _load_data(path=path_sales_per_share)
-    book_value_per_share = _load_data(path=path_book_value_per_share)
 
     # Use the DataFrame for the price and add more data-columns to it.
     df = price_daily
 
     # Only keep the rows where the share-price is defined.
     df.dropna(subset=[SHARE_PRICE], inplace=True)
 
-    # Add financial data to the data-frame (interpolated daily).
-    df[SALES_PER_SHARE] = _resample_daily(sales_per_share)
-    df[BOOK_VALUE_PER_SHARE] = _resample_daily(book_value_per_share)
+    # Load Sales Per Share data.
+    if sales:
+        _load_sales_per_share(ticker=ticker, df=df)
 
-    # Add financial ratios to the data-frame (daily).
-    df[PSALES] = df[SHARE_PRICE] / df[SALES_PER_SHARE]
-    df[PBOOK] = df[SHARE_PRICE] / df[BOOK_VALUE_PER_SHARE]
+    # Load Earnings Per Share data.
+    # This needs the Sales Per Share data to calculate the profit margin.
+    if earnings:
+        _load_earnings_per_share(ticker=ticker, df=df)
+
+    # Load Book-Value Per Share data.
+    if book_value:
+        _load_book_value_per_share(ticker=ticker, df=df)
 
     return df
 

diff --git a/data/CLX Earnings Per Share.txt b/data/CLX Earnings Per Share.txt
@@ -0,0 +1,28 @@
+Date	Earnings Per Share
+06/30/1992	0.45
+06/30/1993	0.76
+06/30/1994	0.99
+06/30/1995	0.95
+06/30/1996	1.07
+06/30/1997	1.21
+06/30/1998	1.46
+06/30/1999	1.05
+06/30/2000	1.67
+06/30/2001	1.37
+06/30/2002	1.39
+06/30/2003	2.26
+06/30/2004	2.31
+06/30/2005	2.93
+06/30/2006	2.94
+06/30/2007	3.31
+06/30/2008	3.30
+06/30/2009	3.86
+06/30/2010	4.30
+06/30/2011	4.07
+06/30/2012	4.13
+06/30/2013	4.36
+06/30/2014	4.31
+06/30/2015	4.45
+06/30/2016	5.00
+06/30/2017	5.44
+
diff --git a/data/CPB Earnings Per Share.txt b/data/CPB Earnings Per Share.txt
@@ -0,0 +1,36 @@
+Date	Earnings Per Share
+07/30/1984	0.37
+07/30/1985	0.38
+07/30/1986	0.43
+07/30/1987	0.48
+07/30/1988	0.53
+07/30/1989	0.03
+07/30/1990	0.01
+07/30/1991	0.79
+07/30/1992	0.97
+07/30/1993	0.02
+07/30/1994	1.26
+07/30/1995	1.40
+07/30/1996	1.61
+07/30/1997	1.51
+07/30/1998	1.45
+07/30/1999	1.64
+07/30/2000	1.68
+07/30/2001	1.57
+07/30/2002	1.28
+07/30/2003	1.45
+07/30/2004	1.58
+07/30/2005	1.73
+07/30/2006	1.88
+07/30/2007	2.21
+07/30/2008	3.12
+07/30/2009	2.09
+07/30/2010	2.48
+07/30/2011	2.47
+07/30/2012	2.44
+07/30/2013	1.46
+07/30/2014	2.61
+07/30/2015	2.21
+07/30/2016	1.82
+07/30/2017	2.91
+
diff --git a/data/DE Earnings Per Share.txt b/data/DE Earnings Per Share.txt
@@ -0,0 +1,31 @@
+Date	Earnings Per Share
+10/29/1989	0.84
+10/29/1990	0.90
+10/29/1991	-0.04
+10/29/1992	0.08
+10/29/1993	-1.99
+10/29/1994	1.17
+10/29/1995	1.36
+10/29/1996	1.57
+10/29/1997	1.89
+10/29/1998	2.10
+10/29/1999	0.51
+10/29/2000	1.04
+10/29/2001	-0.14
+10/29/2002	0.67
+10/29/2003	1.34
+10/29/2004	2.84
+10/29/2005	2.97
+10/29/2006	3.63
+10/29/2007	4.06
+10/29/2008	4.76
+10/29/2009	2.06
+10/29/2010	4.40
+10/29/2011	6.71
+10/29/2012	7.72
+10/29/2013	9.18
+10/29/2014	8.71
+10/29/2015	5.82
+10/29/2016	4.84
+10/29/2017	6.76
+
diff --git a/data/DIS Earnings Per Share.txt b/data/DIS Earnings Per Share.txt
@@ -0,0 +1,26 @@
+Date	Earnings Per Share
+09/30/1994	0.68
+09/30/1995	0.87
+09/30/1996	0.65
+09/30/1997	0.97
+09/30/1998	0.91
+09/30/1999	0.63
+09/30/2000	0.44
+09/30/2001	-0.08
+09/30/2002	0.61
+09/30/2003	0.62
+09/30/2004	1.14
+09/30/2005	1.25
+09/30/2006	1.68
+09/30/2007	2.34
+09/30/2008	2.34
+09/30/2009	1.76
+09/30/2010	2.03
+09/30/2011	2.52
+09/30/2012	3.17
+09/30/2013	3.42
+09/30/2014	4.31
+09/30/2015	4.95
+09/30/2016	5.76
+09/30/2017	5.73
+