Fixed y axis in Python plotting times in 12 hr format

Question

I have this plot but I need the y axis to be fixed to 00:00, 01:00, 02:00, etc all the way up to 12:00. As of now it's only plotting the values I have in the csv on the y axis. the csv is in the following format. How do o get the y axis to be constant and only show 00:00 to 12:00 in 1 hr increments and still have the data plotted correctly?

    ML  INT 0.1     534.15  0:00
    ML  EXT 0.25    654.23  3:00
    ML  INT 0.35    743.12  6:30

And the following is the code I have so far.

    import pandas as pd
    import matplotlib.pyplot as plt
    import numpy as np

    data = pd.read_csv('data.csv', header=None)
    ints = data[data[1]=='INT']
    exts = data[data[1]=='EXT']
    INT_index = data[data[1]=='INT'].index
    EXT_index = data[data[1]=='EXT'].index
    time = [t for t in data[4]]
    int_dist = [d for d in ints[3]]
    ext_dist = [d for d in exts[3]]


    fig, ax = plt.subplots()
    ax.scatter(int_dist, INT_index, c='orange', s=150)
    ax.scatter(ext_dist, EXT_index, c='black', s=150)
    ax.set_yticks(np.arange(len(data[4])))
    ax.set_yticklabels(time)
    plt.legend(['INT', 'EXT'], loc=4)
    plt.xlabel('Distance')
    plt.ylabel('Time')
    plt.show()

Have you tried `ax.set_ylim`? – cphlewis May 14 '15 at 07:53 — cphlewis, May 14 '15 at 07:53
I think I finally solved your problem... – Scott May 14 '15 at 19:32 — Scott, May 14 '15 at 19:32

score 0 · Accepted Answer · answered May 14 '15 at 19:32

I generated a few more rows of data to make the problem, at least on my end, a bit more meaningful.

What solved this for me was generating a 5th column (in code, not the csv) which is the number of minutes corresponding to a particular o'clock time, i.e. 11:59 maps to 719 min. Using pandas I inserted this new column into the dataframe. I could then place string ticklabels for every hour ('0:00', '1:00', etc.) at every 60 min.

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

data = pd.read_csv('Workbook2.csv', header=None)
print data

Prints my faked data:

    0    1     2       3      4
0  ML  INT  0.10  534.15   0:00
1  ML  EXT  0.25  654.23   3:00
2  ML  INT  0.30  743.12   6:30
3  ML  EXT  0.35  744.20   4:30
4  ML  INT  0.45  811.47   7:00
5  ML  EXT  0.55  777.90   5:45
6  ML  INT  0.66  854.70   7:54
7  ML  EXT  0.74  798.40   6:55
8  ML  INT  0.87  947.30  11:59

Now make a function to convert o'clock to minutes:

def convert_to_min(o_clock):
    h, m = o_clock.split(':')
    return int(h) * 60 + int(m)
# using this function create a list times in minutes for each time in col 4
min_col = [convert_to_min(t) for t in data[4]]
data[5] = min_col  # inserts this list as a new column '5'
print data

Our new data:

    0    1     2       3      4    5
0  ML  INT  0.10  534.15   0:00    0
1  ML  EXT  0.25  654.23   3:00  180
2  ML  INT  0.30  743.12   6:30  390
3  ML  EXT  0.35  744.20   4:30  270
4  ML  INT  0.45  811.47   7:00  420
5  ML  EXT  0.55  777.90   5:45  345
6  ML  INT  0.66  854.70   7:54  474
7  ML  EXT  0.74  798.40   6:55  415
8  ML  INT  0.87  947.30  11:59  719

Now build the x and y axis data, the ticklabels, and the tick locations:

INTs = data[data[1]=='INT']
EXTs = data[data[1]=='EXT']

int_dist = INTs[3]  # x-axis data for INT
ext_dist = EXTs[3]

# plotting time as minutes in range [0 720]
int_time = INTs[5]  # y-axis data for INT
ext_time = EXTs[5]

time = ['0:00', '1:00', '2:00', '3:00', '4:00', '5:00', 
        '6:00', '7:00', '8:00', '9:00', '10:00', '11:00', '12:00']
# this will place the strings above at every 60 min
tick_location = [t*60 for t in range(13)]

Now plot:

fig, ax = plt.subplots()
ax.scatter(int_dist, int_time, c='orange', s=150)
ax.scatter(ext_dist, ext_time, c='black', s=150)
ax.set_yticks(tick_location)
ax.set_yticklabels(time)
plt.legend(['INT', 'EXT'], loc=4)
plt.xlabel('Distance')
plt.ylabel('Time')
plt.title('Seems to work...')
plt.show()

enter image description here

yup that did it. thanks it was racking my brain trying to figure it out — Archimedes, May 14 '15 at 21:49
@Archimedes Happy to help. I knew there had to be some simple trick to fix this. Once I thought of converting to minutes it was easy. — Scott, May 14 '15 at 21:51

selwyth · Answer 2 · 2015-05-15T09:16:21.643

The ticks will be a lot smarter if you use a datetime for the y-axis.

Fake data:

df = pd.DataFrame({'value':[530,640,710], 'time':['0:00', '3:00', '6:30']})

   time  value
0  0:00    530
1  3:00    640
2  6:30    710

Convert df.time from str to datetime:

time2 = pd.to_datetime(df.time, format='%H:%M')
plt.plot(df.value, time2, marker='o', linestyle='None')

output

Can't seem to get this into a scatter instead of plot in case it matters for you (I suppressed the line). Maybe because datetime should always be in a timeseries lineplot and never in a scatterplot (I welcome comments that let me know if this is indeed the case and datetime cannot be put into a scatter).

Is there a way for it not to display seconds? – nielsen Jun 09 '20 at 16:50 — nielsen, Jun 09 '20 at 16:50

Fixed y axis in Python plotting times in 12 hr format

2 Answers2

Linked