CODEWITHSUNDEEP
pythonmatplotlibplotgeospatialgeo
ucyo
ucyo

Reputation: 665

Geographical data plot/map with lines in python and matplotlib

I remember seeing on a blog post a nice technique to visualize geographical data. It was just lines representing latitude and the high of the lines the variable to be shown. I tried to sketch it on the following picture: line plot for visualizing geographical data

Does some of you remember the library or even the blog post which explained how to generate these maps? (I vaguely remember it being matplotlib & python, but I could very well be wrong)

Upvotes: 1

Views: 1872

Answers (2)

J Richard Snape
J Richard Snape

Reputation: 20344

I think this is the kind of thing you want - plotting lines of constant latitude on a 3d axis. I've explained what each section does in comments

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import numpy as np
import itertools

#read in data from csv organised in columns labelled 'lat','lon','elevation'
data = np.recfromcsv('elevation-sample.csv', delimiter=',')

# create a 3d axis on a figure
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')

# Find unique (i.e. constant) latitude points
id_list = np.unique(data['lat'])

# stride is how many lines to miss.  set to 1 to get every line
# higher to miss more
stride = 5

# Extract each line from the dataset and plot it on the axes
for id in id_list[::stride]:
    this_line_data = data[np.where(data['lat'] == id)]
    lat,lon,ele = zip(*this_line_data)
    ax.plot(lon,lat,ele, color='black')

# set the viewpoint so we're looking straight at the longitude (x) axis
ax.view_init(elev=45., azim=90)

ax.set_xlabel('Longitude')
ax.set_ylabel('Latitude')
ax.set_zlabel('Elevation')
ax.set_zlim([0,1500])

plt.show()

The data set I used to test is not mine, but I found it on github here.

This gives output as follows:

elevation along constant latitude plot

Note - you can swap latitude and longitude if I've misinterpreted the axis labels in your sketch.

Upvotes: 4

Vidar
Vidar

Reputation: 1877

Are you thinking a 3D plot similar to this? Possibly you could also do a cascade plot like this? The code for the last type of plot is something like this:

# Input parameters:
padding = 1    # Relative distance between plots
ax = gca()     # Matplotlib axes to plot in
spectra = np.random.rand((10, 100))   # Series of Y-data
x_data = np.arange(len(spectra[0]))   # X-data

# Figure out distance between plots:
max_value = 0
for spectrum in spectra:
    spectrum_yrange = (np.nanmax(spectrum) -
                       np.nanmin(spectrum))
    if spectrum_yrange > max_value:
        max_value = spectrum_yrange
# Plot the individual lines
for i, spectrum in enumerate(spectra):
    # Normalize the data to max_value
    data = (spectrum - spectrum.min()) / float(max_value)
    # Offset the individual lines
    data += i * padding
    ax.plot(x_data, data)

Upvotes: 1

Related Questions