stock-correlations-tooltip_calcs.py

def convert_rankings_to_string(ranking):
    """
    Concatenates list of node and correlation into a single string which is the 
    preferred format for the plotly tooltip.

    Inserts html "<br>" inbetween each item in order to add a new line in the tooltip
    """
    s = ''
    for r in ranking:
        s += r + "<br>"
    return s


def calculate_stats(returns=log_returns_df):
    """calculate annualised returns and volatility for all ETFs
   
    Output
    -------
    Outputs the annualised volatility and returns as a list of floats (for use in assigning node colours
    and sizes) and also as a lists of formatted strings to be used in the tool tips.    
    """
    
    #log returns are additive, 252 trading days
    annualized_returns = list(np.mean(returns)*252*100)
    
    annualized_volatility = [np.std(returns[col]*100)*(252**0.5)
                             for col in list(returns.columns)]

   
    # create string for tooltip
    annualized_volatility_2dp = ["Annualized Volatility: ""%.1f" % r + "%" for r in annualized_volatility]
    annualized_returns_2dp = ["Annualized Returns: ""%.1f" % r + "%" for r in annualized_returns]

    return annualized_volatility, annualized_returns, annualized_volatility_2dp, annualized_returns_2dp


def get_top_and_bottom_three(df=correlation_matrix):
    """
    get a list of the top 3 and bottom 3 most/least correlated assests
    for each node. 

    Parameters -> df - pandas correlation matrix

    Returns -> top_3_list : list of lists containing the top 3 correlations (name and value)
               bottom_3_list: list of lists containing the bottom three correlations (name and value)

    """

    top_3_list = []
    bottom_3_list = []

    for col in df.columns:

        # exclude self correlation #reverse order of the list returned
        top_3 = list(np.argsort(abs(df[col]))[-4:-1][::-1])
        # bottom 3 list is returned in correct order
        bottom_3 = list(np.argsort(abs(df[col]))[:3])

        # get column index
        col_index = df.columns.get_loc(col)

        # find values based on index locations
        top_3_values = [df.index[x] + ": %.2f" %
                        df.iloc[x, col_index] for x in top_3]
        bottom_3_values = [df.index[x] + ": %.2f" %
                           df.iloc[x, col_index] for x in bottom_3]

        top_3_list.append(convert_rankings_to_string(top_3_values))
        bottom_3_list.append(convert_rankings_to_string(bottom_3_values))

    return top_3_list, bottom_3_list