绘制图形并在条形图python上包含缺失值数字

Question

python的新手，我确实有医疗数据，其中有基于医院的数据。现在，我确实想绘制每个医院的数据，收集的时间段和患者数，但是我也想在图表中包括每个医院的丢失数据数，这样我也可以知道每个月没有输入数据医院。这是我的数据的摘要

data = {'Hosp_name':['Hos1', 'Hos1', 'Hos2', 'Hos2','Hos3','Hos3'], 'Period':['20-Apr', '21-Apr', '20-Apr', '21-Aug','20-Apr','21-Apr'],
                                                                             'Num_of_patients':[30,'NAN',45,56,'NAN',67]} 
df = pd.DataFrame(data)

如何转换此数据，以便可以正确地绘制它并在图形上包括缺失的数据。这就是我尝试过的

import plotly.express as px
fig = px.bar(df, x='Hosp_name', y=['period','Num_of_patients'])
fig.show()

预期输出是每个医院的条形图，显示了收集的期间数据，缺失值数量和患者数量

Answer 1

这是我的版本，我如何理解以及使用matplotlib会如何

import pandas as pd
data = {'Hosp_name':['Hos1', 'Hos1', 'Hos2', 'Hos2','Hos3','Hos3'], 'Period':['20-Apr', '21-Apr', '20-Apr', '21-Aug','20-Apr','21-Apr'], 'Num_of_patients':[30,'NAN',45,56,'NAN',67]} 
df = pd.DataFrame(data).groupby(['Period','Hosp_name'])[['Num_of_patients']].sum()#.reset_index().set_index('Period')
df['Num_of_patients'] = pd.to_numeric(df['Num_of_patients'], errors = 'coerce')

Out[1]:

                        Num_of_patients
    Period  Hosp_name   
    20-Apr  Hos1       30.0
            Hos2       45.0
            Hos3       NaN
    21-Apr  Hos1       NaN
            Hos3       67.0
    21-Aug  Hos2       56.0

df1 = df.pivot_table(index = 'Period', values = 'Num_of_patients', columns = 'Hosp_name').fillna(0) 
df1

Out[2]:

    Hosp_name   Hos1    Hos2    Hos3
    Period          
    20-Apr      30.0    45.0    0.0
    21-Apr      0.0     0.0    67.0
    21-Aug      0.0     56.0    0.0


df1 = df1.reset_index()

import matplotlib.pyplot as plt
%matplotlib inline

for i,col in enumerate(df1.columns[1:]):
    bars= plt.bar([x+i*0.3 for x in list(df1.index)], df1[col], width = 0.3, label = col)

    for bar in bars:
        plt.gca().text(bar.get_x() + bar.get_width()/2, bar.get_height() - 5, str(int(bar.get_height())),ha='center', color='w', fontsize=11)

plt.xticks(list(df1.index), df1['Period'])
plt.legend()

用于一组列

data = {'Hosp_name':['Hos1', 'Hos1', 'Hos2', 'Hos2','Hos3','Hos3'],'Period':['20-Apr', '21-Apr', '20-Apr', '21-Aug','20-Apr','21-Apr'], 'Num_of_patients':[30,'NAN',45,56,'NAN',67],'age':[11,21,31,36,26,16], 'patient_visits':[20,30,40,50,60,10]} 

# i don't know your case, but i can't imagine the reasin for sum(age), so i evaluate the mean(age)

df = pd.DataFrame(data).groupby(['Period','Hosp_name'])[['Num_of_patients','age','patient_visits']].agg({'Num_of_patients': 'sum', 'patient_visits': 'sum', 'age':'mean'}) 

df['Num_of_patients'] = pd.to_numeric(df['Num_of_patients'], errors = 'coerce')

df
Out[10]:

                   Num_of_patients  patient_visits  age
    Period      Hosp_name           
    20-Apr      Hos1      30.0      20              11
                Hos2      45.0      40              31
                Hos3      NaN       60              26
    21-Apr      Hos1      NaN       30              21 
                Hos3      67.0      10              16
    21-Aug      Hos2      56.0      50              36


df1 = df.pivot_table(index = 'Period', values = 'Num_of_patients','patient_visits','age'], columns = 'Hosp_name').fillna(0)
df1 = df1.reset_index()
df1

Out[11]:

                Period  Num_of_patients         age                   patient_visits
       Hosp_name        Hos1    Hos2    Hos3    Hos1    Hos2    Hos3    Hos1    Hos2    Hos3
           0    20-Apr  30.0    45.0    0.0     11.0    31.0    26.0    20.0    40.0    60.0
           1    21-Apr  0.0     0.0     67.0    21.0    0.0     16.0    30.0    0.0     10.0
           2    21-Aug  0.0     56.0    0.0      0.0    36.0    0.0       0.0   50.0    0.0

这是用于可视化的完整代码（度量标准不同的图）

您可以根据需要进行可视化。

import matplotlib.pyplot as plt
%matplotlib inline

level_1 = ['Num_of_patients', 'age', 'patient_visits']
level_2 = ['Hos1', 'Hos2', 'Hos3']

color = ['r','g','b']

fig, ax = plt.subplots(3,1, sharex = True, gridspec_kw={'hspace': 0.3}, figsize=(10,5))


for i, metric in enumerate(level_1):
    for j, hosp in enumerate(level_2):

        bars = ax[i].bar([x+j*0.2 for x in list(df1.index)], df1[(metric, hosp)], width = 0.2, color = color[j], label = hosp)       

        for bar in bars:
                ax[i].text(bar.get_x() + bar.get_width()/2, bar.get_height() - 8, str(int(bar.get_height())), 
                         ha='center', color='w', fontsize=8)

    ax[i].set_title(f'{metric}', loc = 'right')
    ax[i].legend()


plt.xticks(list(df1.index), df1['Period'])

Hosps的不同图

import matplotlib.pyplot as plt
%matplotlib inline

level_1 = ['Hos1', 'Hos2', 'Hos3']
level_2 = ['Num_of_patients', 'age', 'patient_visits']

color = ['r','g','b']

fig, ax = plt.subplots(3,1, sharex = True, gridspec_kw={'hspace': 0.3}, figsize=(5,5))

for i, hosp in enumerate(level_1):
    for j, metric in enumerate(level_2):

        bars = ax[i].bar([x+j*0.2 for x in list(df1.index)], df1[(metric, hosp)], width = 0.2, color = color[j], label = metric)       

        for bar in bars:
                ax[i].text(bar.get_x() + bar.get_width()/2, bar.get_height() - 8, str(int(bar.get_height())), 
                         ha='center', color='w', fontsize=8)

    ax[i].set_title(f'{metric}', loc = 'right')
    ax[i].legend()


_ = plt.xticks(list(df1.index), df1['Period'])