13_plot_and_select_cluster.py

def generate_plot(clus,i):
    
    plt.style.use('default')

    tags=['Restaurant','Coffee','Food','Pizza','Sandwich']
    colors = []
    for value in clus.index: 
        if  any(t in value for t in tags):
            colors.append('#0000FF')
        else:
            colors.append('#FF0000')

    ax=clus.plot(kind='barh', figsize=(16,8), color=colors, alpha=0.7)

    plt.title('(in % of all venues)\n')
    ax.title.set_fontsize(14)
    plt.suptitle('Ten Most Prevalent Venues of Cluster {}'.format(i), fontsize=16)

    ax.spines['top'].set_visible(False)
    ax.spines['left'].set_visible(False)
    ax.spines['right'].set_visible(False)
    ax.spines['bottom'].set_visible(False)

    plt.xticks([])
    ax.tick_params(axis ='both', which ='both', length = 0)
    labels = [(item.get_text()+'  ') for item in ax.get_yticklabels()]
    ax.set_yticklabels(labels)

    for label in (ax.get_yticklabels()):
        label.set_fontsize(12)

    for index, value in enumerate(clus): 
        label = "%.1f " % round(value*100,1) + "%"
        # place text at the end of bar (adding 0.001 to x, and 0.1 from y to make it appear just after the bar)
        plt.annotate(label, xy=(value + 0.001, index - 0.1), color='black',fontsize=12)

    legend_elements = [Patch(facecolor='#0000FF', edgecolor='#0000FF',
                             label='Food Venues',alpha=0.7),
                       Patch(facecolor='#FF0000', edgecolor='#FF0000',
                             label='Others',alpha=0.7)]

    ax.legend(handles=legend_elements, loc='best',fontsize=12)

    plt.show()
    
    
cluster1=pd.DataFrame(la_results.iloc[1,0:-1]).transpose()
cluster1.sort_values(by='Cluster 1',axis=1,ascending=False,inplace=True)
display(cluster1)

clus1=cluster1.iloc[0,9::-1]
generate_plot(clus1,1)