matplotlib 画图

matplotlib 函数：

饼状图

ax = plt.subplots()
ax.pie()

def pieChart(sdss_df):  # 饼形图
    
    '''绘制饼图以获取标签数量.'''
    label_counts = sdss_df['class'].value_counts()  # DataFrame['Unit Name'].value_counts() 查看表格某列中有多少个不同值，并计算每个不同值有在该列中有多少重复值
    colors = ['skyblue', 'red', 'gold']
    fig1, ax1 = plt.subplots()
    ax1.pie(label_counts, labels=['Galaxy', 'Stars', 'Quasars'], autopct='%1.2f%%', startangle=45, colors=colors) # 画饼状图，并指定标签和对应颜色
    ax1.axis('equal')  # axis equal 将横轴纵轴的定标系数设成相同值
    plt.title('SDSS Object Classes')
    plt.show()
        

---

分布直方图

aseaborn.distplot()
https://zhuanlan.zhihu.com/p/161513797
kdeplot(核密度估计图) https://blog.csdn.net/qq_39949963/article/details/79362501

def distribution(sdss_df, axes, feature, row):
    
    '''Plot the distribution of a space object w.r.t. a given feature.'''
    labels = np.unique(sdss_df['class'])  # np.unique( ) 去除数组中的重复数字,并进行排序之后输出
    colors = ['skyblue', 'gold', 'red']
    
    for i in range(len(labels)):
        label = labels[i]
        # pandas.DataFrame.loc()使用column名和index名进行定位
        # 分布直方图 seaborn.distplot
        ax = sns.distplot(sdss_df.loc[sdss_df['class']==label, feature], kde=False, bins=30, ax=axes[row, i], color=colors[i])
        ax.set_title(label)
        if (i == 0):
            ax.set(ylabel='Count')
        

---

回归图

aseaborn.distplot()
https://zhuanlan.zhihu.com/p/161513797
kdeplot(核密度估计图) https://blog.csdn.net/qq_39949963/article/details/79362501

def distribution(sdss_df, axes, feature, row):
    
    '''Plot the distribution of a space object w.r.t. a given feature.'''
    labels = np.unique(sdss_df['class'])  # np.unique( ) 去除数组中的重复数字,并进行排序之后输出
    colors = ['skyblue', 'gold', 'red']
    
    for i in range(len(labels)):
        label = labels[i]
        # pandas.DataFrame.loc()使用column名和index名进行定位
        # 分布直方图 seaborn.distplot
        ax = sns.distplot(sdss_df.loc[sdss_df['class']==label, feature], kde=False, bins=30, ax=axes[row, i], color=colors[i])
        ax.set_title(label)
        if (i == 0):
            ax.set(ylabel='Count')
        

---

def main():

    # read in SDSS data
    filepath = 'Skyserver_12_30_2019 4_49_58 PM.csv'
    sdss_df = pd.read_csv(filepath, encoding='utf-8')

    # define lists of relevant features
    geo = ['ra', 'dec']  # 经纬度
    nonugriv = ['redshift', 'plate', 'mjd', 'fiberid']
    ugriv = ['u', 'g', 'r', 'i', 'z']

    # 绘制标签计数饼图
    pieChart(sdss_df)

    # 绘制观测的赤道坐标
    for row in range(3):
        equitorial(sdss_df, row)
        plt.show()
    
    # plot the distribution of non-geo and non-ugriv features
    fig, axes = plt.subplots(nrows=4, ncols=3, figsize=(12, 14))
    plt.subplots_adjust(wspace=.4, hspace=.4)
    for row in range(len(nonugriv)):
        feat = nonugriv[row]
        distribution(sdss_df, axes, feat, row)
    plt.show()
        
    # plot the distribution of ugriv features
    fig, axes = plt.subplots(nrows=5, ncols=3, figsize=(12, 15))
    plt.subplots_adjust(wspace=.4, hspace=.4)
    for row in range(len(ugriv)):
        feat = ugriv[row]
        distribution(sdss_df, axes, feat, row)
    plt.show()

main()