Add lines to 3D surface

📊 Plotly Python
1

I am a newbie to Plotly. Is there a way to draw data lines on top of a 3D surface graph. I have managed to draw the graph but cannot figure out how to add lines to it (I have attached the target image). Kindly help. This is my code and dataset.

ch3_ci_trial
ch3_ci_trial750×750 44.7 KB

target
target1020×944 191 KB

#!/usr/bin/env python
#import plotly.graph_objects as go
from plotly.offline import iplot, init_notebook_mode
init_notebook_mode()

import pandas as pd
import copy

z_data = pd.read_csv(‘https://raw.githubusercontent.com/rogomichael/plotpl/master/trial.csv’)
z2_data = pd.read_csv(‘https://raw.githubusercontent.com/rogomichael/plotpl/master/set2.csv’)
#surfacecolor = copy.deepcopy(z2_data)
#surfacecolor[2][2] = 0
fig = go.Figure(data=[
go.Surface(
x = [-5,-4,-3,-2,-1,0,1,2,3,4,5],
y = [-4,-3,-2,-1,0,1,2,3,4,5,6],
z=z2_data,
colorscale = ‘reds’,
opacity = 1,
#colorscale=‘Electric’,
#surfacecolor=surfacecolor,
showscale=False),
go.Surface(
x = [-5,-4,-3,-2,-1,0,1,2,3,4,5],
y = [-4,-3,-2,-1,0,1,2,3,4,5,6],
z=z_data.values,
#colorscale=‘aggrnyl’,
colorscale = ‘greens’,
opacity = 1,
#colorscale = [[0, ‘green’], [-0.5, ‘red’], [-1.0, ‘rgb(0, 0, 255)’]],
showscale=False)])

fig.update_layout(title=‘ch3_ci’,
scene ={“xaxis”:{“title”: “g/NAC [Angs]”},
“yaxis”:{“title”: “h/GD [Angs]”},
“zaxis”:{“title”: “energy/eV”}},
autosize=True,
width=950, height=950,
margin=dict(l=65, r=50, b=65, t=120))

#fig.show()
iplot(fig, image=‘svg’, filename=‘ch3_plot.svg’)

2

Hi @rogomichael,

The Plotly surfaces reveal very well their geometry that’s why I wouldn’t draw wireframes to get them looking like matlab surfaces . There is no trace designed specially for wireframes, but we could define the transversal lines onto the surface as Scatter3d lines:

import numpy as np
import plotly.graph_objects as go

x = np.linspace(-1,1, 50)
y = np.linspace(-1,3, 100)
x, y = np.meshgrid(x,y)

z = x**3-3*y*x+1  # the surface eqn

fig = go.Figure()
fig.add_surface(x=x, y=y, z=z, colorscale='Reds_r', colorbar_thickness=25, colorbar_len=0.75);

X = np.linspace(-1,1, 6)
Y = np.linspace(-1, 3, 12)
#Define the first family of coordinate lines
X, Y = np.meshgrid(X,Y)
Z = X**3-3*Y*X+1
line_marker = dict(color='#101010', width=4)
for xx, yy, zz in zip(X, Y, Z+0.03):
    fig.add_scatter3d(x=xx, y=yy, z=zz, mode='lines', line=line_marker, name='')
#Define the second family of coordinate lines
Y, X = np.meshgrid(Y, X)
Z = X**3-3*Y*X+1
line_marker = dict(color='#101010', width=4)
for xx, yy, zz in zip(X, Y, Z+0.03):
    fig.add_scatter3d(x=xx, y=yy, z=zz, mode='lines', line=line_marker, name='')    
fig.update_layout(width=700, height=700, showlegend=False)
fig.show()
3

Thank you @empet for your answer. Can this be done using z data points from my values or I have to figure out a z equation for this to work in my case. Sorry I am very green in this area. Kindly help .

4

@rogomichael You can use your z-data but, in both cases (surface and lines) z must have the same shape as x, y
in the meshgrid.

5

@empet, I sat down and did a little bit of thinking to produce the result I wanted. Thanks for your suggestions. One more question, how do I shift/push the z values away from the margin/tickslines and how can I do the same with the label without changing the camera because I am interested in the intersecting point view?. I have tried all tricks but none worked
trial_view

6

@rogomichael

To change the surface position with respect to the three coordinate planes I suggest to enlarge the `zaxis_range’
from, let us say, [a, b] to [a-h, b] for some h you’ll choose by trial and error and a similar update for the axis that now is horizontal.

7

@empet Is it possible to define flat face color option (discrete colors for each grid cell) in this surface chart? When I apply surface colors, it is always gradient in nature. Please guide !!

8

@Joemanuel

Could you please post an image with an example, because such a surface is not generated straightforward, and I wouldn’t want to work on something that does not meet your requirements.
Is what you’d like to plot a surface defined over a meshgrid, and the points within the rectangle of vertex coordinates (x,y)
with x, y evaluated at the indices:

    (i, j)      (i, j+1)
    (i+1, j)    (i+1, j+1)

are mapped to the same color?

9

@empet Yes, I am looking for a visualisation similar to the below image.

dd0bgr1r7vw31
dd0bgr1r7vw31852×671 66.7 KB

I am looking for an option to get face colors similar to the above rather than the default gradient pattern. Using some of your previous references, I have build a surface chart for the sphere, scatter3d for the mesh. The current output I get is similar to the below image

mlb

10

This kind of plot seems to be a surface à la Matlab There is a Matlab Plotly version to get something like that: Plotly Graphing Library for MATLAB® | Plotly Graphing Library for MATLAB® | Plotly, Single-Page Reference | Plotly Graphing Library for MATLAB® | Plotly. But since the json file resulted from matlab code is sent to plotly.js, I don’t think it is possible to get such a a plot, because Plotly interpolates the z-values before mapping them to the colorscale.

11

@empet Yes, this is matlab output. Thanks for sharing your inputs, Appreciate it !!!

12

@Joemanuel
I suggest to put your question and the image of your Matlab plot, here: Plotly for MATLABÂŽ - Plotly Community Forum.