I'm trying to plot multiple plots on the same figure using python. The graph should be linear, the x coordinates represent the time of the day, and the y coordinates match the values. Each plot matches a different date.

The data is stored inside a dictionary. The keys represent the dates, and the values hold 2 lists: the first matches the x coordinates, and the second matches the y coordinates. For example:

dict_data = {"4 April": [[datetime(1900, 1, 1, 22, 59), datetime(1900, 1, 1, 23, 1), datetime(1900, 1, 1, 23, 8), datetime(1900, 1, 1, 23, 50)], [405, 320, 300, 360]], "5 April": [[datetime(1900, 1, 1, 8, 10), datetime(1900, 1, 1, 9, 40), datetime(1900, 1, 1, 11, 8), datetime(1900, 1, 1, 11, 10)], [120, 20, 10, 0]]}

I found apost on stack overflow Plotting time in Python with Matplotlib. It was'nt helpful because the x-axis on the graph they created is in "datetime" type, while I use "datetime.time"datetime". (I don't want the x-axis to show the dates). Also, the graph they created is a scatter plot, while I need it to be linear.

This is what I tried:

def multiple_plots(dict_data):
"""
Method to plot multiple times in one figure.
It receives a dictionary with the representation of the data in the csv file.
Every key in the dictionary represent a different date that will have its own plot ont the graph.
"""
for date, coordinates in dict_data.items():
    time_coordinates = coordinates[0]
    # converting the x coordinates in the type datetime.time to int
    x_coordinates = matplotlib.dates.date2num(time_coordinates)
    val_coordinates = coordinates[1]
    plt.plot(list(map(int, x_coordinates)), list(map(int, val_coordinates)), label=date)

plt.legend(loc='best')
plt.show()

CodePudding user response：

This should just work without any fuss:

import matplotlib.pyplot as plt
import datetime

dict_data = {"4 April": [[datetime.datetime(1900, 1, 1, 22, 59), datetime.datetime(1900, 1, 1, 23, 1), datetime.datetime(1900, 1, 1, 23, 8), datetime.datetime(1900, 1, 1, 23, 50)], [405, 320, 300, 360]], "5 April": [[datetime.datetime(1900, 1, 1, 8, 10), datetime.datetime(1900, 1, 1, 9, 40), datetime.datetime(1900, 1, 1, 11, 8), datetime.datetime(1900, 1, 1, 11, 10)], [120, 20, 10, 0]]}

fig, ax = plt.subplots()
for k in dict_data:
    ax.plot(dict_data[k][0], dict_data[k][1])
plt.show()

Obviously that looks a little cramped, but if you use concise converter:

plt.rcParams['date.converter'] = 'concise'

then you get

CodePudding user response：

Edit based on edit in question

The code, using plot_date and line type '-'.

import matplotlib
import numpy as np
from datetime import datetime
import matplotlib.pyplot as plt

dict_data = {"4 April": [[np.datetime64("2022-07-01T22:59:00"), np.datetime64("2022-07-01T23:01:00"), np.datetime64("2022-07-01T23:08:00"), np.datetime64("2022-07-01T23:50:00")], [405, 320, 300, 360]], "5 April": [[np.datetime64("2022-07-01T08:10:00"), np.datetime64("2022-07-01T09:40:00"), np.datetime64("2022-07-01T11:08:00"), np.datetime64("2022-07-01T11:10:00")], [120, 20, 10, 0]]}
# dict_data = {"4 April": [[datetime(2022, 7, 1, 22, 59), datetime(2022, 7, 1, 23, 1), datetime(2022, 7, 1, 23, 8), datetime(2022, 7, 1, 23, 50)], [405, 320, 300, 360]], "5 April": [[datetime(2022, 7, 1, 8, 10), datetime(2022, 7, 1, 9, 40), datetime(2022, 7, 1, 11, 8), datetime(2022, 7, 1, 11, 10)], [120, 20, 10, 0]]}

def multiple_plots(dict_data):
    """
    Method to plot multiple times in one figure.
    It receives a dictionary with the representation of the data in the csv file.
    Every key in the dictionary represent a different date that will have its own plot ont the graph.
    """
    for date, coordinates in dict_data.items():
        time_coordinates = coordinates[0]
        # converting the x coordinates in the type np.datetime64 t2022-07-01To int
        x_coordinates = matplotlib.dates.date2num(time_coordinates)
        val_coordinates = coordinates[1]
        plt.plot_date(list(map(int, x_coordinates)), list(map(int, val_coordinates)), '-', label=date)

    plt.legend(loc='best')
    plt.show()

multiple_plots(dict_data)

Final Output:

The date2num function requires the year, month and date in addition to time. So, two simple fixes are:

1. Sending in year, month and time

dict_data = {"4 April": [[datetime(2022, 7, 1, 22, 59), datetime(2022, 7, 1, 23, 1), datetime(2022, 7, 1, 23, 8), datetime(2022, 7, 1, 23, 50)], [405, 320, 300, 360]], "5 April": [[datetime(2022, 7, 1, 8, 10), datetime(2022, 7, 1, 9, 40), datetime(2022, 7, 1, 11, 8), datetime(2022, 7, 1, 11, 10)], [120, 20, 10, 0]]}

2. Sending in np.datetime64 type

dict_data = {"4 April": [[np.datetime64("2022-07-01T22:59:00"), np.datetime64("2022-07-01T23:01:00"), np.datetime64("2022-07-01T23:08:00"), np.datetime64("2022-07-01T23:50:00")], [405, 320, 300, 360]], "5 April": [[np.datetime64("2022-07-01T08:10:00"), np.datetime64("2022-07-01T09:40:00"), np.datetime64("2022-07-01T11:08:00"), np.datetime64("2022-07-01T11:10:00")], [120, 20, 10, 0]]}

They both work. The final plot is: