有没有一种方法可以在一个 plot 中创建一个堆叠图，或者在一个 plot 中创建带有 matplot/seaborn 的堆叠 Y 轴？

Question

I am new in the visualizations of dataframes and I am exploring the methods, but I do not seem able to find a way to create a very specific plot I need.

I have a dataframe like the one below where I have datetime as index and 4 profiles (whereas there could be more columns uninterested in plotting):

datetime                                                        
2019-06-11 00:00:00  28.97  38.47    NaN  41.47   
2019-06-11 01:00:00  28.83  38.42    NaN  41.48   
2019-06-11 02:00:00  28.72  38.38    NaN  41.49   
2019-06-11 03:00:00  28.56  38.33    NaN  41.49   
2019-06-11 04:00:00  28.36  38.22    NaN  41.51

My aim is to create a plot with the described specifications:

1. Lineplots, one for each profile, but all in one plot .

2. The lines should not intercept, so the profiles should be stacked one above the other, preferably with profile 1 being at the very top.

3. The values should not change, thus the Y-axes should probably be stacked.

4. The legend should be one for the whole plot of the 4 profiles.

5. The plot or line of a profile with NaN values should not appear.

I 'd appreciate it if you could suggest something or guide me through the solution?

Additionally, is there a brief structured guide with examples that you would recommend me in order to gain some good fundamentals in the visualizations? I prefer something I can print out or read via a tablet while commuting.

Thank you for your input:-)

Answer 1

May bi it would be helpful for you:

df[3] = pd.to_numeric(df[3], errors = 'coerce').fillna(0)
df = df.rename(columns = {1:'A',2:'B',3:'C',4:'D'})
df = df.set_index(0)
df

Out[1]:

                        A      B    C    D
                 0              
2019-06-11 00:00:00 28.97   38.47   0.0 41.47
2019-06-11 01:00:00 28.83   38.42   0.0 41.48
2019-06-11 02:00:00 28.72   38.38   0.0 41.49
2019-06-11 03:00:00 28.56   38.33   0.0 41.49
2019-06-11 04:00:00 28.36   38.22   0.0 41.51


import matplotlib.dates as mdates


ax = [None]*4

fig = plt.figure()
f, (ax[0], ax[1], ax[2], ax[3]) = plt.subplots(4, 1, sharex=True)

ax[3].set_xlabel('Hours')

color = ['r','g','b','y']
for i in range(4):
    col = df.columns[i]
    _ = ax[i].plot(df.index,df[col], label = col, c = color[i])

xfmt = mdates.DateFormatter('%H:%M')
ax[3].xaxis.set_major_formatter(xfmt)
ax[3].xaxis.set_major_locator(mdates.HourLocator(byhour=range(24)))


_ = f.legend()