cat_card_by_cluster_plot.py

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

from math import pi

import matplotlib as mpl
plt.rcParams["font.family"] = 'NanumGothicCoding'
mpl.rcParams['axes.unicode_minus'] = False

def cat_card_by_cluster_plot(cluster: int, data, save=False):
    plt.figure(figsize=(10,10))
    ax = plt.subplot(polar=True)

    # set categories
    categories = data.category.unique().tolist()
    
    N = len(categories)
    
    # set color
    color=['red','blue','orange','green','purple']

    # calculate values
    for i in range(5):
        df2019 = data[data['t'] <= '2019-06-30']
        df2020 = data[('2020-01-01' <= data['t']) & (data['t'] <= '2020-06-30')]
        
        df2019_card_mean = df2019[df2019.k == i].groupby('category')['card'].mean()
        df2020_card_mean = df2020[df2020.k == i].groupby('category')['card'].mean()
        

        values = (df2020_card_mean / df2019_card_mean).tolist()
        values += values[:1]

        # angles
        angles = [n / float(N) * 2 * pi for n in range(N)]
        angles += angles[:1]
        
        if i == cluster:
            plt.polar(angles, values, marker='.', linewidth=2, color=color[i])
            plt.fill(angles, values, alpha=0.5, color=color[i])
        else:
            plt.polar(angles, values, marker='.', linewidth=1, linestyle='--', color=color[i], alpha=0.5)
                
        plt.xticks(angles[:-1], categories)
        
        ax.spines["polar"].set_visible(False)
        
        yticks_values = [0.2, 0.4, 0.6, 0.8, 1]
        plt.yticks(yticks_values, labels=['{0:.0%}'.format(i) for i in yticks_values], 
                    color='grey', size=17)
        plt.xticks(size=20)
    
    plt.title('군집별 업종별 구매 건수 변화율', size=25)
    plt.legend([f'군집{i}' for i in range(5)], fontsize=20, bbox_to_anchor=(1.45, 0.8))
    plt.tight_layout()
    
    if save:
        plt.savefig('../images/sample/card_example.jpg', dpi=300)
    
    plt.show()
   

categories = ['업종1','업종2','업종3','업종4','업종5','업종6']
sex_age = [f'{sex}_{age}' for sex in ['F','M'] for age in range(10,80,10)]
date = pd.date_range('2019-01-01','2020-06-30',freq='d')

sample_df = pd.DataFrame()
for sex_age_i in sex_age:
    for cat in categories:
        sample_i_df = pd.DataFrame({'t':date})
        sample_i_df['sex_age'] = sex_age_i
        sample_i_df['category'] = cat

        sample_df = pd.concat([sample_df, sample_i_df], axis=0)

sample_df = sample_df.sample(frac=1, random_state=42)

np.random.seed(42)
sample_df['card'] = random_walks(n_ts=1, sz=len(sample_df), d=1)[0,:,0]
sample_df['card'] = sample_df['card'] - np.random.randint(low=1, high=100, size=len(sample_df))
sample_df['k'] = np.random.randint(low=0, high=5, size=len(sample_df))

cat_card_by_cluster_plot(cluster=3, data=sample_df, save=False)