Matplotlib 图形 x 刻度是在所有 x 数据点之后

Question

I am graphing timestamps against integers on a binary step graph, but it had all the timestamps on the x axis that are plotted instead of intervals, so I tried manually setting the x ticks to every 15 minutes throughout a day, but all the ticks seem to be after the data points:

Without trying to manually set the ticks:

The data that is being plotted is in the following format:

[['21:20:24', '21:21:15', '51', '1'], ['21:27:55', '21:29:15', '80', '1'], ['21:38:16', '21:40:15', '119', '1'], ['22:13:44', '22:19:31', '347', '3'], ['22:24:42', '22:26:15', '93', '1'], ['22:47:29', '22:49:15', '106', '1']] (which is produced by self.padDs() function)

What can I do to fix the ticks?

#Plot as binary step graph
def binaryGraph(self, day):
    fig = plot.figure(figsize=(17, 2))
    ax = fig.add_subplot(111)

    allData = self.getAllFromDir("datastream/"+day+"/")
    i = 1
    for ds in allData:
        padded = self.padDs(ds)
        x = []
        y = []
        for line in padded:
            x.append(line[0])
            y.append(int(line[2]))
            x.append(line[1])
            y.append(int(line[2]))

        ax.step(x,y, label=day+str(i))
        i = i + 1
        break

    plot.title(day)
    plot.xlabel("Time")
    plot.ylabel("Interval length (s)")
    plot.legend()
    ticks = []
    for i in range(0, 25):
        tick = str(i)+":"
        if i < 10:
            tick = "0"+str(i)+":"

        for j in range(0, 60, 15):
            mins = tick + str(j)
            if j == 0:
                mins = tick + "0"+ str(j)
            ticks.append(mins)

    plot.xticks(ticks)
    print(ticks)
    plot.show()

Answer 1

Calling plt.xticks with just one parameter of strings isn't a good approach. The first parameter to xticks should be a list of numbers, places in the x-axis where you want a tick. The optional second parameter would be the list of strings to be written at those ticks.

In your case, you have just strings for the x-axis. Internally, they are just numbers 0,1,2,... . This means that in the current plot, the width of all the steps is equal, being exactly one position.

One way to go further, is to make the x-axis numerically, for example measuring in minutes. That way the width of your steps can be proportional to the interval length.

In that case, you can calculate ticks for every 15 minutes. And tick labels for each of these ticks as hours and minutes ( HH:MM ).

Here is some code to do just that. I chose a smaller font and rotated the labels to make them fit on the page.

from matplotlib import pyplot as plt

fig = plt.figure(figsize=(17, 2))
ax = fig.add_subplot(111)

padded = [['21:20:24', '21:21:15', '51', '1'], ['21:27:55', '21:29:15', '80', '1'], ['21:38:16', '21:40:15', '119', '1'], ['22:13:44', '22:19:31', '347', '3'], ['22:24:42', '22:26:15', '93', '1'], ['22:47:29', '22:49:15', '106', '1']]
x = []
y = []
for line in padded:
    # change x's from text to a number in minutes
    x0, x1 = [int(h) * 60 + int(m) + int(s) / 60.0 for i in (0, 1) for h, m, s in [line[i].split(':')]]
    x.append(x0)
    y.append(int(line[2]))
    x.append(x1)
    y.append(int(line[2]))

ax.step(x, y, label="day_i")

plt.title("day")
plt.xlabel("Time")
plt.ylabel("Interval length (s)")
plt.legend()
ticks = []
labels = []
for i in range(0, 25):
    for j in range(0, 60, 15):
        labels.append(f"{i:02}:{j:02}")
        ticks.append(i * 60 + j)
        if i == 24:  # make 24:00 the last tick
            break
plt.xticks(ticks, labels, rotation=90, fontsize=8)
plt.tight_layout()
plt.show()

The code to generate the ticks could be made more Pythonic as follows.

ticks = range(0, 24*60+1, 15)
labels = [f"{m // 60:02}:{m % 60:02}" for m in ticks]

The advantage is that you can just change the ticks and the labels will still be OK. For example:

ticks = range(0, 24*60+1, 60)
labels = [f"{m // 60:02}:{m % 60:02}" for m in ticks]
minor_ticks = range(0, 24*60+1, 15)
plt.xticks(ticks, labels, fontsize=10)
ax.xaxis.set_ticks(minor_ticks, minor=True)

Probably, while creating the y array, you could set the second y to 0 , as the step plot already takes care of drawing steps, and the 0 would be useful for the inbetween-intervals without data. In that case, you need to add where='post' to the step plot (to indicate that a new step is valid for after the x0). And you'd need to add an extra 0 at the start and end of the day ( ax.step([0]+x+[24*60], [0]+y+[0], where='post', label=...) ).

Zooming in on your example data would look like: