how to format the y axis for a timedelta object in plotly express

Question

I’ve seen several related questions, but none of the solutions so far seem to solve my problem. The problem is, that instead of e.g. "03:00:00" or similar I get 40T as label on the y axis for my timedelta object. It is nicely formatted in pandas, though: 0 days 03:00:00

The output format is either unusable, e.g. when converting to unsorted strings, or the output doesn’t change at all.

I would like to have an easily readable format on the y axis instead of the seconds (?), it will most often be durations of hours and minutes, in a few cases it might be longer than a day (but something like 40:xx:xx meaning 40 hours would be totally fine

import pandas as pd
data = [['tom', 10, "2023-06-21 06:23:55+00:00", "2023-06-21 09:23:55+00:00"], ['nick', 15, "2023-06-20 06:23:55+00:00", "2023-06-21 06:23:55+00:00"], ['juli', 14, "2023-06-21 06:23:50+00:00", "2023-06-21 06:23:55+00:00"]]
df = pd.DataFrame(data, columns=['name', 'age', "start", "stop"])
df["start"] = pd.to_datetime(df["start"])
df["stop"] = pd.to_datetime(df["stop"])
df["duration"] = df["stop"] - df["start"]
df["duration"]

#### output
0   0 days 03:00:00
1   1 days 00:00:00
2   0 days 00:00:05
Name: duration, dtype: timedelta64[ns]

import plotly.express as px
import plotly.graph_objects as go

fig = px.scatter(
        df,
        x=df["age"].sort_values(ascending=False),
        y="duration", # I guess it shows the seconds
        #y=pd.to_timedelta(df.duration, unit='h'), # same format as before
        #y=df["duration"].sort_values(ascending=True).dt.to_pytimedelta().astype(str), # fixed label with equal distance between marks, regardless of numerical difference
        color="name",
    )
figure = go.Figure(data=fig)
# figure.update_layout(yaxis_tickformat='%H:%M:%S') # adds a lot of zeroes?
# figure.update_layout(yaxis_tickformat="%H:%M:%S.%f")
figure.show()

Asked By: crazysantaclaus

||

Source

Answer 1

Plotly does not seem to handle time-delta format as time-series data; the plotly community is discussing how to handle time-delta. In this context, the time format can be applied by giving the base date of the time series to the time delta value as a workaround.
However, this workaround should be used with the understanding that it is not valid for all time delta values.

import plotly.express as px
import plotly.graph_objects as go

fig = px.scatter(
        df,
        x=df["age"].sort_values(ascending=False),
        y=df["duration"] + pd.to_datetime('1970/01/01'), 
        color="name",
    )
figure = go.Figure(data=fig)
figure.update_layout(yaxis_tickformat="%H:%M:%S.%f")
figure.show()

Answered By: r-beginners

how to format the y axis for a timedelta object in plotly express

Question:

Answers: