Plot time periods in timeline from pandas

Question

I don't know if this is possible to do with matplotlib or other visualization tools in Anaconda distribution.

I know how to do this with JavaScript and may just go that route, but figured I'd see if there is a way to do it in my Jupyter notebook. I have multiple time periods with a start and end datetime for each row of data I want to plot. My goal is actually to find gaps where there is no time period in any row.

Here is an example: http://visjs.org/examples/timeline/other/groupsPerformance.html

All rows have at least one time period, but some have multiple periods. Is there a way to do this in matplotlib or seaborn?

A pointer to specific documentation for this kind of visualization or an example is sufficient.

Sample dataframe:

    id           start1        end1      start2        end2
0  Bob       2018-11-29  2018-11-30  2018-12-01  2018-12-31
1  James     2018-10-19  2018-10-31         NaT         NaT
2  Jane      2018-04-05  2018-07-12  2018-11-29  2018-11-30

So the Y-axis would contain the id field, and X-axis would be the time. Each row would have (broken) horizontal bars for each , where applicable (there may or may not be a second period).

Recently I had to implement a similar plot using matplotlib. I used `barh` but, unfortunately, I had to do most of the calculations for the plot myself.. [just like this](https://stackoverflow.com/questions/43367690/how-to-get-gantt-plot-using-matplotlib) — wendelbsilva, Nov 30 '18 at 18:10

score 2 · Answer 1 · answered Nov 30 '18 at 18:27

2

broken_barh() in matplotlib would seem to fit your needs. I'd provide more detail, but for that I'd need some sample data.

answered Nov 30 '18 at 18:27

ASGM

11,051
1
32
53

I appreciate it! – Arthur Dent Nov 30 '18 at 20:19
I added sample data FYI if you know of an easy way - otherwise will try to figure it out – Arthur Dent Nov 30 '18 at 20:36

Vivek Kalyanarangan · Accepted Answer · 2018-11-30T21:39:37.530

2

If you are open to using plotly, this can take you closer -

df_full = df[['id','start1','end1']].rename(columns={'id':'Task','start1':'Start','end1':'Finish'}).append(
    df[['id','start2','end2']].rename(columns={'id':'Task','start2':'Start','end2':'Finish'}),sort=False).reset_index()

import plotly.plotly as py
import plotly.figure_factory as ff

fig = ff.create_gantt(df_full)
py.iplot(fig, filename='gantt-simple-gantt-chart', world_readable=True)

EDIT

To group tasks together, use -

fig = ff.create_gantt(df_full, group_tasks=True)
py.iplot(fig, filename='gantt-group-tasks-together', world_readable=True)

edited Nov 30 '18 at 21:39

answered Nov 30 '18 at 20:47

Vivek Kalyanarangan

8,951
1
23
42

1

This looks pretty good, although ideally only one line per row - if possible. But still very good start – Arthur Dent Nov 30 '18 at 21:01

score 0 · Answer 3 · answered Jan 09 '21 at 08:47

0

you can actually integrate a javascript based timeline in Jupyter Notebook (with all the flexibility it gives). I wrote an explanation here.

answered Jan 09 '21 at 08:47

Alon

353
1
10

Plot time periods in timeline from pandas

3 Answers3