制作分布图类似密度图,在python中利用pandas来提取分布数据是比较方便的。主要用到pandas的cut和groupby等函数。
import pandas
from sqlalchemy import create_engine
host_mysql_test = '127.0.0.1'
port_mysql_test = 3306
user_mysql_test = 'admin'
pwd_mysql_test = '1234'
db_name_mysql_test = 'mydb'
engine_hq = create_engine('mysql+mysqldb://%s:%s@%s:%d/%s' % (user_mysql_test,
pwd_mysql_test,
host_mysql_test,
port_mysql_test,
'hq_db'), connect_args={'charset': 'utf8'})
sql = "SELECT * FROM fund_data where quarter>=8 order by yanzhi desc"
df = pd.read_sql(sql, engine)
#将yanzhi数据转换为百分比
df['yanzhi'] = df['yanzhi'].apply(lambda x: x * 100)
pandas.cut(x, bins, right=True, labels=None, retbins=False, precision=3, include_lowest=False)
官方文档链接
主要参数为x和bins。
x为数据源,数组格式的都支持,list,numpy.narray, pandas.Series。
bins可以为int,也可以为序列。
bins = [0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
cats = pd.cut(df['yanzhi'], bins)
我们定义bins为一个序列,默认为左开右闭的区间:
In[]:print cats
Out[]:
0 (90, 100]
1 (90, 100]
2 (90, 100]
3 (80, 90]
4 (80, 90]
...
970 (10, 20]
971 (10, 20]
972 (10, 20]
973 (10, 20]
974 (10, 20]
Name: yanzhi, dtype: category
Categories (10, object): [(0, 10], (10, 20], (20, 30], (30, 40], ..., (60, 70], (70, 80], (80, 90]
, (90, 100]]
对言值列按cats做groupby,然后调用get_stats统计函数,再用unstack函数将层次化的行索引“展开”为列。
def get_stats(group):
return {'count': group.count()}
grouped = df['yanzhi'].groupby(cats)
bin_counts = grouped.apply(get_stats).unstack()
print bin_counts
count
yanzhi
(0, 10] 0
(10, 20] 5
(20, 30] 22
(30, 40] 92
(40, 50] 258
(50, 60] 357
(60, 70] 178
(70, 80] 51
(80, 90] 9
(90, 100] 3
bin_counts.index = ['0~10', '10~20', '20~30', '30~40', '40~50', '50~60', '60~70',
'70~80', '80~90', '90~100']
bin_counts.index.name = 'yanzhi'
bin_counts.plot(kind='bar', alpha=0.5, rot=0)
G2在之前的文章中有介绍,文章《python结合G2绘制精美图形》。
datas = []
for ix, row in bin_counts.iterrows():
# if row['机构数量'] > 0:
sss = {'name': ix, 'count': row['count']}
datas.append(sss)
encodejson = json.dumps(datas, ensure_ascii=False)
f = open('yanzhi.json', 'w')
f.write(encodejson)
f.close()
分布图
一句话绘制出来,但具体的区间段难以区分出来。
df["yanzhi"].hist(bins=20, alpha=0.5)
bokeh是python的一个优秀的绘图工具包,与pandas结合的比较好。bokeh文档
from bokeh.charts import Histogram, output_file,show
hist=Histogram(df, values='yanzhi',bins=30, title='分布图', legend='top_right')
output_file('hist.html', title='hist example')
show(hist)
作者原文链接:python制作分布图