[FIXED] How to show the minor ticks labels in matplotlib

Issue

I am trying to plot a pandas dataframe, but I can't show the labels on the minor ticks

The dataframe is something like this:

day_of_week   hour   count
0             0      150
0             1      673
...
1             0      734
1             1      35
...

where day_of_week is the day of the week (Monday, Tuesday, etc) and hour are the hours of the day from 0 to 23.

I want to plot this quantity as a time-series and show the major ticks as mid-night and change of the next day and the minor ticks every 6 hours.

I modified a bit the dataframe and tried to plot it, this is the code:

days = {0: "Mon", 1: "Tue", 2: "Wed", 3: "Thu", 4: "Fri", 5: "Sat", 6: "Sun"}

df["dow"] = df["day_of_week"].apply(lambda x: days.get(x))
df["axis"] = (df["dow"] + " " + df["hour"].astype(str))

fig, ax = plt.subplots()

ax.plot(time_tweets["axis"], time_tweets["count"])

ax.grid(which="major", axis="x", linestyle="-", linewidth=2)

ax.xaxis.set_major_locator(MultipleLocator(24))
ax.xaxis.set_minor_locator(MultipleLocator(6))

plt.xticks(rotation=45, minor=True, ha="right")
plt.xticks(rotation=45, minor=False, ha="right")

This doesn't plot the label on the minor ticks, everything else is fine, the major ticks and label, the minor ticks and the grid only on the major ticks.

I tried to add plt.minorticks_on() or plt.setp(ax.get_xticklabels(minor=True), visible=True), but nothing label on the minor ticks.

Do you have any suggestion?

Solution

If you create fake dates, you can use DayLocator and HourLocator:

base = pd.to_datetime('2023-12-18')  # an arbitrary Monday
offset = pd.to_timedelta(df['day_of_week'], unit='D') \
         + pd.to_timedelta(df['hour'], unit='H')

df['dti'] = base + offset
print(df)

# Output
     day_of_week  hour  count                 dti
0              0     0    584 2023-12-18 00:00:00
1              0     1     63 2023-12-18 01:00:00
2              0     2    368 2023-12-18 02:00:00
3              0     3    493 2023-12-18 03:00:00
4              0     4    864 2023-12-18 04:00:00
..           ...   ...    ...                 ...
163            6    19     42 2023-12-24 19:00:00
164            6    20    960 2023-12-24 20:00:00
165            6    21      5 2023-12-24 21:00:00
166            6    22    440 2023-12-24 22:00:00
167            6    23    295 2023-12-24 23:00:00

[168 rows x 4 columns]

import matplotlib.dates as mdates

fig, ax = plt.subplots(figsize=(12, 8))

ax.plot(df['dti'], df['count'])

ax.grid(which="major", axis="x", linestyle="-", linewidth=2)

ax.xaxis.set_major_locator(mdates.DayLocator())
ax.xaxis.set_minor_locator(mdates.HourLocator([6, 12, 18]))
ax.xaxis.set_major_formatter(mdates.DateFormatter('%a'))
ax.xaxis.set_minor_formatter(mdates.DateFormatter('%H'))

plt.xticks(rotation=45, minor=True, ha="right")
plt.xticks(rotation=45, minor=False, ha="right")
plt.show()

Output:

Answered By - Corralien