Anaconda install instructions as of mgwr 1.0.2 and Conda 4.5:

conda update conda
conda update anaconda
conda update scipy
conda install pysal 
pip install pysal.lib
conda install mgwr
conda install geopandas
import numpy as np
import pysal.lib as ps 
from pysal.model.mgwr.gwr import GWR, MGWR
from pysal.model.mgwr.sel_bw import Sel_BW
from pysal.model.mgwr.utils import shift_colormap, truncate_colormap
import geopandas as gp
import matplotlib.pyplot as plt
import matplotlib as mpl
import pandas as pd

#Load Georgia dataset and generate plot of Georgia counties (figure 1)
georgia_data = pd.read_csv(ps.examples.get_path('GData_utm.csv'))
georgia_shp = gp.read_file(ps.examples.get_path('G_utm.shp'))
fig, ax = plt.subplots(figsize=(10,10))
georgia_shp.plot(ax=ax, **{'edgecolor':'black', 'facecolor':'white'})
georgia_shp.centroid.plot(ax=ax, c='black')

<matplotlib.axes._subplots.AxesSubplot at 0x1a18c0b6a0>

png

#Prepare Georgia dataset inputs
g_y = georgia_data['PctBach'].values.reshape((-1,1))
g_X = georgia_data[['PctFB', 'PctBlack', 'PctRural']].values
u = georgia_data['X']
v = georgia_data['Y']
g_coords = list(zip(u,v))

g_X = (g_X - g_X.mean(axis=0)) / g_X.std(axis=0)

g_y = g_y.reshape((-1,1))

g_y = (g_y - g_y.mean(axis=0)) / g_y.std(axis=0)

#Calibrate GWR model

gwr_selector = Sel_BW(g_coords, g_y, g_X)
gwr_bw = gwr_selector.search(bw_min=2)
print(gwr_bw)
gwr_results = GWR(g_coords, g_y, g_X, gwr_bw).fit()

117.0
#Calibrate MGWR model

mgwr_selector = Sel_BW(g_coords, g_y, g_X, multi=True)
mgwr_bw = mgwr_selector.search(multi_bw_min=[2])
print(mgwr_bw)
mgwr_results = MGWR(g_coords, g_y, g_X, mgwr_selector).fit()

[ 92. 101. 136. 158.]
mgwr_results.summary()

===========================================================================
Model type                                                         Gaussian
Number of observations:                                                 159
Number of covariates:                                                     4

Global Regression Results
---------------------------------------------------------------------------
Residual sum of squares:                                             71.793
Log-likelihood:                                                    -162.399
AIC:                                                                332.798
AICc:                                                               335.191
BIC:                                                               -713.887
R2:                                                                   0.548
Adj. R2:                                                              0.540

Variable                              Est.         SE  t(Est/SE)    p-value
------------------------------- ---------- ---------- ---------- ----------
X0                                   0.000      0.054      0.000      1.000
X1                                   0.458      0.066      6.988      0.000
X2                                  -0.084      0.055     -1.525      0.127
X3                                  -0.374      0.065     -5.734      0.000

Multi-Scale Geographically Weighted Regression (MGWR) Results
---------------------------------------------------------------------------
Spatial kernel:                                           Adaptive bisquare
Criterion for optimal bandwidth:                                       AICc
Score of Change (SOC) type:                                     Smoothing f
Termination criterion for MGWR:                                       1e-05

MGWR bandwidths
---------------------------------------------------------------------------
Variable             Bandwidth      ENP_j   Adj t-val(95%)   Adj alpha(95%)
X0                      92.000      3.845            2.512            0.013
X1                     101.000      3.514            2.479            0.014
X2                     136.000      2.258            2.311            0.022
X3                     158.000      1.752            2.210            0.029

Diagnostic information
---------------------------------------------------------------------------
Residual sum of squares:                                             50.899
Effective number of parameters (trace(S)):                           11.368
Degree of freedom (n - trace(S)):                                   147.632
Sigma estimate:                                                       0.587
Log-likelihood:                                                    -135.056
AIC:                                                                294.849
AICc:                                                               297.120
BIC:                                                                332.806
R2                                                                    0.680
Adjusted R2                                                           0.655

Summary Statistics For MGWR Parameter Estimates
---------------------------------------------------------------------------
Variable                   Mean        STD        Min     Median        Max
-------------------- ---------- ---------- ---------- ---------- ----------
X0                        0.017      0.171     -0.260      0.058      0.271
X1                        0.479      0.216      0.117      0.500      0.722
X2                       -0.069      0.036     -0.146     -0.064     -0.014
X3                       -0.304      0.019     -0.347     -0.302     -0.266
===========================================================================

#Prepare GWR results for mapping

#Add GWR parameters to GeoDataframe
georgia_shp['gwr_intercept'] = gwr_results.params[:,0]
georgia_shp['gwr_fb'] = gwr_results.params[:,1]
georgia_shp['gwr_aa'] = gwr_results.params[:,2]
georgia_shp['gwr_rural'] = gwr_results.params[:,3]

#Obtain t-vals filtered based on multiple testing correction
gwr_filtered_t = gwr_results.filter_tvals()

#Prepare MGWR results for mapping

#Add MGWR parameters to GeoDataframe
georgia_shp['mgwr_intercept'] = mgwr_results.params[:,0]
georgia_shp['mgwr_fb'] = mgwr_results.params[:,1]
georgia_shp['mgwr_aa'] = mgwr_results.params[:,2]
georgia_shp['mgwr_rural'] = mgwr_results.params[:,3]

#Obtain t-vals filtered based on multiple testing correction
mgwr_filtered_t = mgwr_results.filter_tvals()

#Comparison maps of GWR vs. MGWR parameter surfaces where the grey units pertain to statistically insignificant parameters

#Prep plot and add axes
fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(45,20))
ax0 = axes[0]
ax0.set_title('GWR Intercept Surface (BW: ' + str(gwr_bw) +')', fontsize=40)
ax1 = axes[1]
ax1.set_title('MGWR Intercept Surface (BW: ' + str(mgwr_bw[0]) +')', fontsize=40)

#Set color map
cmap = plt.cm.seismic

#Find min and max values of the two combined datasets
gwr_min = georgia_shp['gwr_intercept'].min()
gwr_max = georgia_shp['gwr_intercept'].max()
mgwr_min = georgia_shp['mgwr_intercept'].min()
mgwr_max = georgia_shp['mgwr_intercept'].max()
vmin = np.min([gwr_min, mgwr_min])
vmax = np.max([gwr_max, mgwr_max])

#If all values are negative use the negative half of the colormap
if (vmin < 0) & (vmax < 0):
    cmap = truncate_colormap(cmap, 0.0, 0.5)
#If all values are positive use the positive half of the colormap
elif (vmin > 0) & (vmax > 0):
    cmap = truncate_colormap(cmap, 0.5, 1.0)
#Otherwise, there are positive and negative values so the colormap so zero is the midpoint
else:
    cmap = shift_colormap(cmap, start=0.0, midpoint=1 - vmax/(vmax + abs(vmin)), stop=1.)

#Create scalar mappable for colorbar and stretch colormap across range of data values
sm = plt.cm.ScalarMappable(cmap=cmap, norm=plt.Normalize(vmin=vmin, vmax=vmax))

#Plot GWR parameters
georgia_shp.plot('gwr_intercept', cmap=sm.cmap, ax=ax0, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
#If there are insignificnt parameters plot gray polygons over them
if (gwr_filtered_t[:,0] == 0).any():
    georgia_shp[gwr_filtered_t[:,0] == 0].plot(color='lightgrey', ax=ax0, **{'edgecolor':'black'})

#Plot MGWR parameters
georgia_shp.plot('mgwr_intercept', cmap=sm.cmap, ax=ax1, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
#If there are insignificnt parameters plot gray polygons over them
if (mgwr_filtered_t[:,0] == 0).any():
    georgia_shp[mgwr_filtered_t[:,0] == 0].plot(color='lightgrey', ax=ax1, **{'edgecolor':'black'})
 
#Set figure options and plot 
fig.tight_layout()    
fig.subplots_adjust(right=0.9)
cax = fig.add_axes([0.92, 0.14, 0.03, 0.75])
sm._A = []
cbar = fig.colorbar(sm, cax=cax)
cbar.ax.tick_params(labelsize=50) 
ax0.get_xaxis().set_visible(False)
ax0.get_yaxis().set_visible(False)
ax1.get_xaxis().set_visible(False)
ax1.get_yaxis().set_visible(False)
plt.show()

png

fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(45,20))
ax0 = axes[0]
ax0.set_title('GWR Percent Foreign Born Surface (BW: ' + str(gwr_bw) +')', fontsize=40)
ax1 = axes[1]
ax1.set_title('MGWR Percent Foreign Born Surface (BW: ' + str(mgwr_bw[1]) +')', fontsize=40)
cmap = plt.cm.seismic
gwr_min = georgia_shp['gwr_fb'].min()
gwr_max = georgia_shp['gwr_fb'].max()
mgwr_min = georgia_shp['mgwr_fb'].min()
mgwr_max = georgia_shp['mgwr_fb'].max()
vmin = np.min([gwr_min, mgwr_min])
vmax = np.max([gwr_max, mgwr_max])

if (vmin < 0) & (vmax < 0):
    cmap = truncate_colormap(cmap, 0.0, 0.5)
elif (vmin > 0) & (vmax > 0):
    cmap = truncate_colormap(cmap, 0.5, 1.0)

cmap = shift_colormap(cmap, start=0.0, midpoint=1 - vmax/(vmax + abs(vmin)), stop=1.)
    
sm = plt.cm.ScalarMappable(cmap=cmap, norm=plt.Normalize(vmin=vmin, vmax=vmax))

georgia_shp.plot('gwr_fb', cmap=sm.cmap, ax=ax0, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
if (gwr_filtered_t[:,1] == 0).any():
    georgia_shp[gwr_filtered_t[:,1] == 0].plot(color='lightgrey', ax=ax0, **{'edgecolor':'black'})

georgia_shp.plot('mgwr_fb', cmap=sm.cmap, ax=ax1, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
if (mgwr_filtered_t[:,1] == 0).any():
    georgia_shp[mgwr_filtered_t[:,1] == 0].plot(color='lightgrey', ax=ax1, **{'edgecolor':'black'})
    
fig.tight_layout()    
fig.subplots_adjust(right=0.9)
cax = fig.add_axes([0.92, 0.14, 0.03, 0.75])
sm._A = []
cbar = fig.colorbar(sm, cax=cax)
cbar.ax.tick_params(labelsize=50) 
ax0.get_xaxis().set_visible(False)
ax0.get_yaxis().set_visible(False)
ax1.get_xaxis().set_visible(False)
ax1.get_yaxis().set_visible(False)
plt.show()

png

fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(45,20))
ax0 = axes[0]
ax0.set_title('GWR Percent African American Surface (BW: ' + str(gwr_bw) +')', fontsize=40)
ax1 = axes[1]
ax1.set_title('MGWR Percent African American Surface (BW: ' + str(mgwr_bw[2]) +')', fontsize=40)
cmap = plt.cm.seismic
gwr_min = georgia_shp['gwr_aa'].min()
gwr_max = georgia_shp['gwr_aa'].max()
mgwr_min = georgia_shp['mgwr_aa'].min()
mgwr_max = georgia_shp['mgwr_aa'].max()
vmin = np.min([gwr_min, mgwr_min])
vmax = np.max([gwr_max, mgwr_max])

if (vmin < 0) & (vmax < 0):
    cmap = truncate_colormap(cmap, 0.0, 0.5)
elif (vmin > 0) & (vmax > 0):
    cmap = truncate_colormap(cmap, 0.5, 1.0)
else:
    cmap = shift_colormap(cmap, start=0.0, midpoint=1 - vmax/(vmax + abs(vmin)), stop=1.)
    
sm = plt.cm.ScalarMappable(cmap=cmap, norm=plt.Normalize(vmin=vmin, vmax=vmax))

georgia_shp.plot('gwr_aa', cmap=sm.cmap, ax=ax0, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
if (gwr_filtered_t[:,2] == 0).any():
    georgia_shp[gwr_filtered_t[:,2] == 0].plot(color='lightgrey', ax=ax0, **{'edgecolor':'black'})

georgia_shp.plot('mgwr_aa', cmap=sm.cmap, ax=ax1, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
if (mgwr_filtered_t[:,2] == 0).any():
    georgia_shp[mgwr_filtered_t[:,2] == 0].plot(color='lightgrey', ax=ax1, **{'edgecolor':'black'})
    
fig.tight_layout()    
fig.subplots_adjust(right=0.9)
cax = fig.add_axes([0.92, 0.14, 0.03, 0.75])
sm._A = []
cbar = fig.colorbar(sm, cax=cax)
cbar.ax.tick_params(labelsize=50) 
ax0.get_xaxis().set_visible(False)
ax0.get_yaxis().set_visible(False)
ax1.get_xaxis().set_visible(False)
ax1.get_yaxis().set_visible(False)
plt.show()

png

fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(45,20))
ax0 = axes[0]
ax0.set_title('GWR Percent Rural Surface (BW: ' + str(gwr_bw) +')', fontsize=40)
ax1 = axes[1]
ax1.set_title('MGWR Percent Rural Surface (BW: ' + str(mgwr_bw[3]) +')', fontsize=40)
cmap = plt.cm.seismic
gwr_min = georgia_shp['gwr_rural'].min()
gwr_max = georgia_shp['gwr_rural'].max()
mgwr_min = georgia_shp['mgwr_rural'].min()
mgwr_max = georgia_shp['mgwr_rural'].max()
vmin = np.min([gwr_min, mgwr_min])
vmax = np.max([gwr_max, mgwr_max])

if (vmin < 0) & (vmax < 0):
    cmap = truncate_colormap(cmap, 0.0, 0.5)
elif (vmin > 0) & (vmax > 0):
    cmap = truncate_colormap(cmap, 0.5, 1.0)
else:
    cmap = shift_colormap(cmap, start=0.0, midpoint=1 - vmax/(vmax + abs(vmin)), stop=1.)
    
sm = plt.cm.ScalarMappable(cmap=cmap, norm=plt.Normalize(vmin=vmin, vmax=vmax))

georgia_shp.plot('gwr_rural', cmap=sm.cmap, ax=ax0, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
if (gwr_filtered_t[:,3] == 0).any():
    georgia_shp[gwr_filtered_t[:,3] == 0].plot(color='lightgrey', ax=ax0, **{'edgecolor':'black'})

georgia_shp.plot('mgwr_rural', cmap=sm.cmap, ax=ax1, vmin=vmin, vmax=vmax, **{'edgecolor':'black', 'alpha':.65})
if (mgwr_filtered_t[:,3] == 0).any():
    georgia_shp[mgwr_filtered_t[:,3] == 0].plot(color='lightgrey', ax=ax1, **{'edgecolor':'black'})

fig.tight_layout()    
fig.subplots_adjust(right=0.9)
cax = fig.add_axes([0.92, 0.14, 0.03, 0.75])
sm._A = []
cbar = fig.colorbar(sm, cax=cax)
cbar.ax.tick_params(labelsize=50) 
ax0.get_xaxis().set_visible(False)
ax0.get_yaxis().set_visible(False)
ax1.get_xaxis().set_visible(False)
ax1.get_yaxis().set_visible(False)
plt.show()

png