pyplot画散点图

画图能够使我们直观的分析数据的特点，图形画的清晰易懂能使我们更好的找到数据的特征，发现其中的规律。
参考链接
https://matplotlib.org/users/pyplot_tutorial.html plot()函数的官方文档
https://www.zealseeker.com/archives/matplotlib-legend-and-text-label/ 讲解了图例和标注
http://note4code.com/2015/03/30/%E4%BD%BF%E7%94%A8matplotlib%E7%BB%98%E5%88%B6%E6%95%A3%E7%82%B9%E5%9B%BE/ 散点图样例
https://github.com/Phlya/adjustText/blob/master/examples/Examples.ipynb 关于标注的详细例子，非常非常好

1.使用scatter()画散点图

import matplotlib.pyplot as pltimport numpy as npn = 100plt.figure(figsize=(7,7))for color in ['red','blue','green']:    x, y = np.random.rand(2, n)    scale = 100*np.random.rand(n)    #s 表示散点的大小，形如 shape (n, )    #label 表示显示在图例中的标注    #alpha 是 RGBA 颜色的透明分量    #edgecolors 指定三点圆周的颜色        plt.scatter(x,y,c=color,s=scale,label=color,alpha=0.6,edgecolors='white')plt.title('Scatter')plt.xlabel('x')plt.ylabel('y')plt.legend(loc='best')plt.grid(True)plt.show()

结果如图

2.使用plot()函数实现
数据资源请访问链接http://download.csdn.net/download/zhousishuo/9973961

import pandas as pdimport numpy as npfrom sklearn.cluster import KMeansimport matplotlib.pyplot as pltdata = pd.read_csv("xfs.csv")df_xfs = data.sort_values(['brand_code', 'red_price']).drop("brand_code",axis=1)#只选取价格不大于150的df_xfs = df_xfs[df_xfs.red_price<=150]#这段代码是为了找出聚类价格的中心点，通过查看有几个销量峰值来确定#销量前十五top15_qtty = df_xfs.drop("red_price",axis=1).sort_values("sale_qtty",ascending=False).head(n=15)#前十五销量的均值avg_qtty = top15_qtty["sale_qtty"].mean()#找到大于前15销量均值的个数center_num = top15_qtty[top15_qtty.sale_qtty>avg_qtty].count()["sale_qtty"]fig = plt.figure(figsize=(15,5))ax1 = fig.add_subplot(1,1,1)#参数ms是markersize的缩写，来控制圆的大小。alpha是RGBA 颜色的透明分量。mec是markeredgecolor缩写，指定圆周颜色。 ax1.plot(df_xfs["red_price"], df_xfs["sale_qtty"],'ro',label='xfs',alpha=0.6,ms=6,mec='white')#画k-means找到的聚类中心点xfsMatrix = df_xfs.drop("sale_qtty",axis=1).as_matrix()xfs_kmeans = KMeans(n_clusters=center_num, random_state=0).fit(xfsMatrix)# print(xfs_kmeans.cluster_centers_)xfs_pd = pd.DataFrame(xfs_kmeans.cluster_centers_, columns = ['red_price'])#这是一个二维的坐标轴，如果不加"sale_qtty"这一列，会在同一坐标轴x上重建坐标，达不到效果xfs_pd["sale_qtty"] = 0# print(xfs_pd)ax1.plot(xfs_pd["red_price"],xfs_pd["sale_qtty"],'go',label='price_centers')#画出前center_num的销量点df_topn = df_xfs.sort_values("sale_qtty",ascending=False).head(n=center_num)topname = 'top' + str(center_num)ax1.plot(df_topn["red_price"],df_topn["sale_qtty"],'bs',label=topname)ax1.legend(loc='best')#为X轴设置一个名称ax1.set_xlabel("price")#为Y轴设置一个名称ax1.set_ylabel("sale_qtty")#设置一个标题ax1.set_title('the relationships of price and sale_qtty')#构造销量前centers_num的坐标x_price = df_topn["red_price"].as_matrix()y_qtty = df_topn["sale_qtty"].as_matrix()text = []texts = []for i in range(len(x_price)):    text.append("("+str(x_price[i])+", "+str(y_qtty[i])+")")for x, y, s in zip(x_price, y_qtty, text):    texts.append(plt.text(x, y, s))plt.show()

结果如图