How to base color line segment on value of a string/ other column?

Question

dataframe with 3 columns

x       y     type
21/02   5     'a'
22/02   6     'b'
...   ...   ...

in total i have two type: 'a' and 'b'

based on a stackoverflow question and matplotlib documentation i came with following:

fig, ax = plt.subplots()   
xy = np.column_stack((df['x'],df['y'])) 
xy = xy.reshape(-1, 1, 2) 
segments = np.hstack([xy[:-1], xy[1:]]) 
coll = LineCollection(segments, color='r') 
ax.add_collection(coll)
plt.show()

This gives me following graph:

I think I have to do something with: coll.set_array(some_value) But I don't know how.

I got this far: I used same code as above but just changed one line:

coll = LineCollection(segments,cmap=plt.cm.gist_ncar)

But now the color of the segments are at random.

How do I base the color of the segments on the column 'type' in my dataframe?

answer for this question is a simple if statement.. you can do it anyway you like. just compare the df["type'] — Eshaka, Feb 11 '19 at 07:13
if you want the colors to be random try using a loop to plot one by one instead of plotting everything together — Eshaka, Feb 11 '19 at 07:15

T. Kau · Answer 1 · 2019-02-11T07:15:59.553

2

If you just have two colors, this should work:

colors = ["red" if type=="a" else "blue" for type in df["type"]] 
coll = LineCollection(segments, color=colors)

For more colors you can map the types to a list of colors in a similar way.

edited Feb 11 '19 at 07:15

answered Feb 11 '19 at 07:07

T. Kau

593
1
4
17

1

do you mean: `coll = LineCollection(segments, color=colors)` ? – J.A.Cado Feb 11 '19 at 07:10
1

color = 'r' will always keep it red – Eshaka Feb 11 '19 at 07:11
Yes, my bad :D Edited it. – T. Kau Feb 11 '19 at 07:16
1

Bam! Works like a charm. Didn’t know you could pass a list with colors with same length as number of lines in collection. Tnx – J.A.Cado Feb 12 '19 at 04:21

How to base color line segment on value of a string/ other column?

1 Answers1