Plotting in 3D with plotly

As plotly offers interactive plots, it becomes feasible to create a 3D plot and rotate it using mouse-interaction.

from datetime import datetime
import pandas as pd
import plotly.express as px

df = pd.read_csv("data/blobs_measurements.csv")
df.head()

	Unnamed: 0	label	area	mean_intensity	max_intensity	min_intensity	centroid-0	centroid-1	eccentricity	solidity	perimeter	centroid_x	centroid_y
0	0	1	27.0	149.333333	168.0	128.0	26.259259	0.666667	0.973991	0.900000	20.621320	0.666667	26.259259
1	1	2	203.0	199.960591	240.0	128.0	225.876847	3.955665	0.892348	0.966667	59.591883	3.955665	225.876847
2	2	3	503.0	198.695825	248.0	128.0	97.487078	7.836978	0.801503	0.971042	86.870058	7.836978	97.487078
3	3	4	264.0	189.939394	232.0	120.0	139.545455	7.159091	0.605384	0.977778	58.041631	7.159091	139.545455
4	4	5	96.0	166.416667	200.0	128.0	44.812500	7.833333	0.628475	0.960000	34.142136	7.833333	44.812500

The scatter_3d function offers interactive 3D plots.

fig = px.scatter_3d(df, 
                    x="area", 
                    y="perimeter", 
                    z="solidity") 
fig.update_traces(marker_size = 2)  # Make dots smaller
fig.update_layout(margin=dict(l=0, r=0, b=0, t=0)) # Reduce figure margins
fig.show()

Exercise

Configure the plot above so that eccentricity is colour-coded in the plot.