finsec_fund_analysis_step06.py

def sns_heatmap(dataframe, key_column, key_column_format, chart_title, c_pos_neg):
    # Build out heatmap things
    dataframe['Position'] = range(1,len(dataframe) + 1)
    dataframe['y'] = [(x//10 + 1 if x%10 != 0 else (x//10)) for x in dataframe['Position']]
    dataframe['x'] = [(x%10 if x%10 != 0 else 10) for x in dataframe['Position']]
    # most_bought['normalized_dollars_bought_sold'] = (most_bought['dollars_bought_sold'] - most_bought['dollars_bought_sold'].min()) / (most_bought['dollars_bought_sold'].max()-most_bought['dollars_bought_sold'].min())
    # most_bought['natural_log'] = np.log(most_bought['normalized_dollars_bought_sold'])
    pivot_df = dataframe.pivot(index='y', columns='x', values=key_column)
    
    ticker_labels = np.asarray(dataframe['Ticker']).reshape((10,10))
    key_column_labels = np.asarray(dataframe[key_column]).reshape((10,10))
    if key_column_format == '$M':
        annot_labels = np.asarray(["{} \n ${:,.0f}M".format(ticker, dbs/1000000) for ticker, dbs in zip(ticker_labels.flatten(), key_column_labels.flatten())]).reshape((10,10))
    elif key_column_format == '$B':
        annot_labels = np.asarray(["{} \n ${:,.0f}B".format(ticker, dbs/1000000000) for ticker, dbs in zip(ticker_labels.flatten(), key_column_labels.flatten())]).reshape((10,10))
    elif key_column_format == '#T':
        annot_labels = np.asarray(["{} \n {:,.0f} Thousand".format(ticker, dbs/1000) for ticker, dbs in zip(ticker_labels.flatten(), key_column_labels.flatten())]).reshape((10,10))
    elif key_column_format == '%':
        annot_labels = np.asarray(["{} \n {:.2f} %".format(ticker, kcv) for ticker, kcv in zip(ticker_labels.flatten(), key_column_labels.flatten())]).reshape((10,10))
    else:
        annot_labels = np.asarray(["{}".format(ticker) for ticker in ticker_labels.flatten()]).reshape((10,10))
    
    fig, ax = plt.subplots(1, 1, figsize = (50, 50), dpi=300)

    cmap_val = 'BuGn' if c_pos_neg == 'pos' else 'OrRd_r'


    sns.heatmap(pivot_df, annot=annot_labels, annot_kws={"size":5}, fmt="", cmap=cmap_val, linewidths=0.25, square=False, yticklabels=False, xticklabels=False, ax=ax)
    # sns.set(font_scale=1.8)
    ax.set_ylabel('')
    ax.set_xlabel('')
    ax.set_title(chart_title)
    plt.show()
    return

# --- Build out heatmaps ---
most_bought = ticker_dataframe[:100].reindex()
sns_heatmap(most_bought, 'dollars_bought_sold', '$M', 'Most Bought Asset by Dollar Value', c_pos_neg='pos')

most_sold = ticker_dataframe[-100:].reindex()
sns_heatmap(most_sold, 'dollars_bought_sold', '$M', 'Most Sold Asset by Dollar Value', c_pos_neg='neg')

most_bought_sold_by_allocation = ticker_dataframe.sort_values(by=['avg_bought_sold_weight_by_portfolio_pct'], ascending=False)[:100]
sns_heatmap(most_bought_sold_by_allocation, 'avg_bought_sold_weight_by_portfolio_pct', '%', 'Most Bought Asset by Portfolio Allocation', c_pos_neg='pos')