Timeline bar graph using python and matplotlib

Question

I am looking to draw a timeline bar graph using matplotlib that will show the things a person did in one day. I am adding the code below's output and an expected output that I am looking for. Any library can be used, in my case the closest I could get to was using matplotlib. Any help would be greatly appreciated.

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

data = [    (dt.datetime(2018, 7, 17, 0, 15), dt.datetime(2018, 7, 17, 0, 30), 'sleep'),
            (dt.datetime(2018, 7, 17, 0, 30), dt.datetime(2018, 7, 17, 0, 45), 'eat'),
            (dt.datetime(2018, 7, 17, 0, 45), dt.datetime(2018, 7, 17, 1, 0), 'work'),
            (dt.datetime(2018, 7, 17, 1, 0), dt.datetime(2018, 7, 17, 1, 30), 'sleep'),
            (dt.datetime(2018, 7, 17, 1, 15), dt.datetime(2018, 7, 17, 1, 30), 'eat'), 
            (dt.datetime(2018, 7, 17, 1, 30), dt.datetime(2018, 7, 17, 1, 45), 'work')
        ]

rng=[]
for i in range(len(data)):
    rng.append((data[i][0]).strftime('%H:%M'))

index={}
activity = []
for i in range(len(data)):
    index[(data[i][2])]=[]
    activity.append(data[i][2])

for i in range(len(index)):
    for j in range(len(activity)):
        if activity[j]==index.keys()[i]:
            index[index.keys()[i]].append(15)
        else:
            index[index.keys()[i]].append(0)            

data = list(index.values())
df = pd.DataFrame(data,index=list(index.keys()))
df.plot.barh(stacked=True, sharex=False)
plt.show()

My Output:

Using matplotlib this is what I was getting

Expected Output:

I got this using google charts' Timeline graph but I need this using python and the data used for generating both graphs is not exactly the same, I hope you get the point

there's an issue with your code @ index[(data[i][3])]=[]

Yuca
– Yuca

2018-07-24 18:50:28 +00:00
Commented Jul 24, 2018 at 18:50 — Yuca
– Yuca, Commented Jul 24, 2018 at 18:50
That's a long lunch

jtlz2
– jtlz2

2022-03-30 17:27:15 +00:00
Commented Mar 30, 2022 at 17:27 — jtlz2
– jtlz2, Commented Mar 30, 2022 at 17:27

ImportanceOfBeingErnest · Accepted Answer · 2018-07-24 19:27:20Z

41

You may create a PolyCollection of "bars". For this you would need to convert your dates to numbers (matplotlib.dates.date2num).

import datetime as dt
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
from matplotlib.collections import PolyCollection

data = [    (dt.datetime(2018, 7, 17, 0, 15), dt.datetime(2018, 7, 17, 0, 30), 'sleep'),
            (dt.datetime(2018, 7, 17, 0, 30), dt.datetime(2018, 7, 17, 0, 45), 'eat'),
            (dt.datetime(2018, 7, 17, 0, 45), dt.datetime(2018, 7, 17, 1, 0), 'work'),
            (dt.datetime(2018, 7, 17, 1, 0), dt.datetime(2018, 7, 17, 1, 30), 'sleep'),
            (dt.datetime(2018, 7, 17, 1, 15), dt.datetime(2018, 7, 17, 1, 30), 'eat'), 
            (dt.datetime(2018, 7, 17, 1, 30), dt.datetime(2018, 7, 17, 1, 45), 'work')
        ]

cats = {"sleep" : 1, "eat" : 2, "work" : 3}
colormapping = {"sleep" : "C0", "eat" : "C1", "work" : "C2"}

verts = []
colors = []
for d in data:
    v =  [(mdates.date2num(d[0]), cats[d[2]]-.4),
          (mdates.date2num(d[0]), cats[d[2]]+.4),
          (mdates.date2num(d[1]), cats[d[2]]+.4),
          (mdates.date2num(d[1]), cats[d[2]]-.4),
          (mdates.date2num(d[0]), cats[d[2]]-.4)]
    verts.append(v)
    colors.append(colormapping[d[2]])

bars = PolyCollection(verts, facecolors=colors)

fig, ax = plt.subplots()
ax.add_collection(bars)
ax.autoscale()
loc = mdates.MinuteLocator(byminute=[0,15,30,45])
ax.xaxis.set_major_locator(loc)
ax.xaxis.set_major_formatter(mdates.AutoDateFormatter(loc))

ax.set_yticks([1,2,3])
ax.set_yticklabels(["sleep", "eat", "work"])
plt.show()

Note that such plots can equally be generated with a Broken Bar plot (broken_barh), however, the (unsorted) data used here, make it a bit easier using a PolyCollection.

And now you would need to explain to me how you can sleep and eat at the same time - something I can never quite get at, as hard as I try.

edited Jul 24, 2018 at 19:27

answered Jul 24, 2018 at 19:14

ImportanceOfBeingErnest

342k61 gold badges737 silver badges771 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

Moritz Over a year ago

Haha, when I saw that a new answer was added, I immediately knew that it was you.

m12345 Over a year ago

Thank you for the answer, that was an error by me in data that you pointed out at the end of your answer, thats the reason i added the expected output in case there was any mistake in my code or question

m12345 Over a year ago

so can we change the x-axis something like 1,2,3..24 hrs or maybe 2,4..24 like in the above example output and plot the graph against it so that it is not clumsy when there is more data throughout the day?

ImportanceOfBeingErnest Over a year ago

Sure, you may use any Locator and Formatter you like. Possibly an HourLocator makes sense and a DateFormatter("%H").

Martin Over a year ago

Sleep eating is a favorite past time for me as well.

|

f3lix · Accepted Answer · 2019-12-05 11:30:40Z

My solution using Altair (example):

import altair as alt
import datetime as dt
import pandas as pd


alt.renderers.enable('jupyterlab')

data = pd.DataFrame()
data['from'] = [dt.datetime(2018, 7, 17, 0, 15),
             dt.datetime(2018, 7, 17, 0, 30),
             dt.datetime(2018, 7, 17, 0, 45), 
             dt.datetime(2018, 7, 17, 1, 0), 
             dt.datetime(2018, 7, 17, 1, 15), 
             dt.datetime(2018, 7, 17, 1, 30)]
data['to'] = [dt.datetime(2018, 7, 17, 0, 30),
             dt.datetime(2018, 7, 17, 0, 45),
             dt.datetime(2018, 7, 17, 1, 0), 
             dt.datetime(2018, 7, 17, 1, 15), 
             dt.datetime(2018, 7, 17, 1, 30), 
             dt.datetime(2018, 7, 17, 1, 45)]
data['activity'] = ['sleep','eat','work','sleep','eat','work']
#data

alt.Chart(data).mark_bar().encode(
    x='from',
    x2='to',
    y='activity',
    color=alt.Color('activity', scale=alt.Scale(scheme='dark2'))
)

Output:

Collectives™ on Stack Overflow

Timeline bar graph using python and matplotlib

2 Answers 2

7 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

7 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related