def crosscorrs(self, clusters=None, bin_width=0.0015, limit=0.03,
figsize=(9,5)):
""" Plots of cross-correlations of clustered spike times. """
times = self.clusters.select(clusters).times()
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
# Set the number of rows and columns to plot
n_rows, n_cols = [len(times)]*2
fig, axes = plt.generate_crosscorr_axes(n_rows, n_cols, num=4,
figsize=figsize)
for (ii, jj) in axes:
ax = axes[(ii,jj)]
cl1, cl2 = times.keys()[ii], times.keys()[jj]
t1, t2 = times[cl1], times[cl2]
# Get the cross-correlation for different clusters
if ii != jj:
plt.crosscorr(t1, t2, ax=ax, bin_width=bin_width, limit=limit)
ax.set_xticklabels('')
ax.set_yticklabels('')
# If cluster 1 is the same as cluster 2, get the autocorrelation
else:
plt.autocorr(t1, ax=ax, color=colors[cl1],
bin_width=bin_width, limit=limit)
ax.set_ylabel('{}'.format(cl1))
ax.set_xticklabels('')
ax.set_yticklabels('')
return fig
def spikes(self, clusters=None, limit=50, figsize=None):
""" Generates plots of clustered spike waveforms.
"""
cls = self.clusters.select(clusters)
cl_spikes = cls.spikes()
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
fig, axes = plt.generate_axes(len(cls),
4,
num=2,
sharex=True,
figsize=figsize)
for ax, cl in zip(axes, cl_spikes):
ax.clear()
spks = cl_spikes[cl]
plt.spikes(spks,
ax=ax,
color=colors[cl],
patch_size=spks.shape[1] / 4)
ax.set_title('Cluster {}'.format(cl))
ax.set_ylabel('Voltage (uV)')
ax.set_xticklabels('')
fig.tight_layout()
return fig
def autocorrs(self, clusters=None, bin_width=0.0015, limit=0.03,
figsize=None):
""" Plots of autocorrelations of clustered spike times.
**Keywords**:
*clusters*: list or iterable
List of clusters to plot
*bin_width*: float
Width of bins in the autocorrelation calculation
*limit*: float
Time limit over which to calculate the autocorrelation
"""
cls = self.clusters.select(clusters)
cl_times = cls.times()
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
fig, axes = plt.generate_axes(len(cls), 4, num=3, figsize=figsize,
sharex=True)
for ax, cl in zip(axes, cl_times):
ax.clear()
tstamps = cl_times[cl]
plt.autocorr(tstamps, ax=ax, color=colors[cl],
bin_width=bin_width, limit=limit)
ax.set_title('Cluster {}'.format(cl))
ax.set_xlabel('Lag (ms)')
fig.tight_layout()
return fig
def plot_top_countries_by_crops(self, grp_cnt, col_name='', xlabel=''):
"""
Plot top countries based on how much global ag area they occupy
:param grp_cnt:
:param col_name:
:param xlabel:
:return:
"""
logger.info('plot_top_countries_by_crops')
cols = plots.get_colors()
# Number of countries to plot is given by constants.PLOT_CNTRS
ax = grp_cnt[col_name][:constants.PLOT_CNTRS].plot(kind='barh',
color=cols[0],
linewidth=0)
# Remove spines from top and right of plot
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)
# Remove y-axis label
ax.set_ylabel('')
ax.set_xlabel(xlabel)
# Ensure that the axis ticks only show up on the bottom and left of the plot
ax.get_xaxis().tick_bottom()
ax.get_yaxis().tick_left()
# Final layout adjustment and output
plt.tight_layout()
plt.savefig(constants.out_dir + 'cntrs_by_area.png', dpi=constants.DPI)
plt.close()
def plot_top_countries_by_crops(self, grp_cnt, col_name='', xlabel=''):
"""
Plot top countries based on how much global ag area they occupy
:param grp_cnt:
:param col_name:
:param xlabel:
:return:
"""
logger.info('plot_top_countries_by_crops')
cols = plots.get_colors()
# Number of countries to plot is given by constants.PLOT_CNTRS
ax = grp_cnt[col_name][:constants.PLOT_CNTRS].plot(kind='barh', color=cols[0], linewidth=0)
# Remove spines from top and right of plot
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)
# Remove y-axis label
ax.set_ylabel('')
ax.set_xlabel(xlabel)
# Ensure that the axis ticks only show up on the bottom and left of the plot
ax.get_xaxis().tick_bottom()
ax.get_yaxis().tick_left()
# Final layout adjustment and output
plt.tight_layout()
plt.savefig(constants.out_dir + 'cntrs_by_area.png', dpi=constants.DPI)
plt.close()
def _scatter_helper(self, clusters=None, limit=500):
""" A helper method to generate the data for the scatter plots. """
cls = self.clusters.select(clusters)
# Here, limit the data so that we don't plot everything
cls = Clusters({cl:plt.limit_data(cls[cl], limit) for cl in cls})
# Get the color and feature arrays for fast plotting
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
col_array = plt.color_array(cls, colors)
feats = cls.features().flatten()
return feats, col_array
def _scatter_helper(self, clusters=None, limit=500):
""" A helper method to generate the data for the scatter plots. """
cls = self.clusters.select(clusters)
# Here, limit the data so that we don't plot everything
cls = Clusters({cl: plt.limit_data(cls[cl], limit) for cl in cls})
# Get the color and feature arrays for fast plotting
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
col_array = plt.color_array(cls, colors)
feats = cls.features().flatten()
return feats, col_array
def plot_cnt_decade(self, inp_fao_df, cnt, already_processed=False):
"""
Plot percentage of cropland area occupied by each crop functional type for a country
:param inp_fao_df:
:param cnt:
:param already_processed:
:return:
"""
out_dec_df = self.per_CFT_by_decade(inp_fao_df, cnt, already_processed)
out_dec_df = out_dec_df.set_index(self.cft_type)
ax = out_dec_df.drop(self.name_country_col, axis=1).T.\
plot(kind='bar', stacked=True, color=plots.get_colors(palette='tableau'), linewidth=0)
plots.simple_axis(ax) # Simple axis, no axis on top and right of plot
# Transparent legend in lower left corner
leg = plt.legend(loc='lower left', fancybox=None)
leg.get_frame().set_linewidth(0.0)
leg.get_frame().set_alpha(0.5)
# Set X and Y axis labels and title
ax.set_title(cnt)
ax.set_xlabel('')
plt.ylim(ymax=100)
ax.set_ylabel(
'Percentage of cropland area \noccupied by each crop functional type'
)
fmt = '%.0f%%' # Format you want the ticks, e.g. '40%'
yticks = mtick.FormatStrFormatter(fmt)
ax.yaxis.set_major_formatter(yticks)
# remove ticks from X axis
plt.tick_params(
axis='x', # changes apply to the x-axis
which='both', # both major and minor ticks are affected
bottom='off', # ticks along the bottom edge are off
top='off') # ticks along the top edge are off
# Rotate the X axis labels to be horizontal
locs, labels = plt.xticks()
plt.setp(labels, rotation=0)
plt.tight_layout()
plt.savefig(constants.out_dir + os.sep + cnt + '.png',
bbox_inches='tight',
dpi=600)
plt.close()
def plot_stacked_country_cft(self,
df,
arr_legend,
path_out,
fname,
ncols=2,
xlabel='',
ylabel='',
title=''):
"""
:param df:
:param arr_legend:
:param path_out:
:param fname:
:param xlabel:
:param ylabel:
:param title:
:return:
"""
df = (df.reset_index().fillna(0.0))
ax = (df.plot.bar(stacked=True,
colormap=plots.get_colors(palette='tableau',
cmap=True),
linewidth=0,
use_index=False))
# Legend
leg = ax.legend(fancybox=None, ncol=ncols, prop={'size': 6})
leg.get_frame().set_linewidth(0.0)
leg.get_frame().set_alpha(0.5)
# Create nice-looking grid for ease of visualization
ax.grid(which='minor', alpha=0.2, linestyle='--')
ax.grid(which='major', alpha=0.5, linestyle='--')
plt.xticks(df.index, df[self.FAO_code])
plt.ylabel(ylabel)
plt.xlabel(xlabel)
plt.title(title)
plt.tight_layout()
plt.savefig(path_out + os.sep + fname, dpi=constants.DPI)
plt.close()
def plot_cnt_mean_decade(self, inp_fao_df, cnt, already_processed=False):
"""
Plot mean crop functional type area in each decade
:param inp_fao_df:
:param cnt:
:param already_processed:
:return:
"""
out_dec_df = self.per_CFT_by_decade(inp_fao_df, cnt, already_processed)
out_dec_df = out_dec_df.set_index(self.cft_type)
ax = out_dec_df.drop(self.name_country_col, axis=1).T.\
plot(kind='bar', stacked=True, color=plots.get_colors(palette='tableau'), linewidth=0)
plots.simple_axis(ax) # Simple axis, no axis on top and right of plot
# Transparent legend in lower left corner
leg = plt.legend(loc='lower left', fancybox=None)
leg.get_frame().set_linewidth(0.0)
leg.get_frame().set_alpha(0.5)
# Set X and Y axis labels and title
ax.set_title(cnt)
ax.set_xlabel('')
plt.ylim(ymax=100)
ax.set_ylabel('Mean crop functional type area in each decade')
fmt = '%.0f%%' # Format you want the ticks, e.g. '40%'
yticks = mtick.FormatStrFormatter(fmt)
ax.yaxis.set_major_formatter(yticks)
# remove ticks from X axis
plt.tick_params(
axis='x', # changes apply to the x-axis
which='both', # both major and minor ticks are affected
bottom='off', # ticks along the bottom edge are off
top='off') # ticks along the top edge are off
# Rotate the X axis labels to be horizontal
locs, labels = plt.xticks()
plt.setp(labels, rotation=0)
plt.tight_layout()
plt.savefig(constants.out_dir + os.sep + 'Mean_' + cnt + '.png', bbox_inches='tight', dpi=600)
plt.close()
def spikes(self, clusters=None, limit=50, figsize=None):
""" Generates plots of clustered spike waveforms.
"""
cls = self.clusters.select(clusters)
cl_spikes = cls.spikes()
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
fig, axes = plt.generate_axes(len(cls), 4, num=2, sharex=True,
figsize=figsize)
for ax, cl in zip(axes, cl_spikes):
ax.clear()
spks = cl_spikes[cl]
plt.spikes(spks, ax=ax, color=colors[cl], patch_size=spks.shape[1]/4)
ax.set_title('Cluster {}'.format(cl))
ax.set_ylabel('Voltage (uV)')
ax.set_xticklabels('')
fig.tight_layout()
return fig
def crosscorrs(self,
clusters=None,
bin_width=0.0015,
limit=0.03,
figsize=(9, 5)):
""" Plots of cross-correlations of clustered spike times. """
times = self.clusters.select(clusters).times()
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
# Set the number of rows and columns to plot
n_rows, n_cols = [len(times)] * 2
fig, axes = plt.generate_crosscorr_axes(n_rows,
n_cols,
num=4,
figsize=figsize)
for (ii, jj) in axes:
ax = axes[(ii, jj)]
cl1, cl2 = times.keys()[ii], times.keys()[jj]
t1, t2 = times[cl1], times[cl2]
# Get the cross-correlation for different clusters
if ii != jj:
plt.crosscorr(t1, t2, ax=ax, bin_width=bin_width, limit=limit)
ax.set_xticklabels('')
ax.set_yticklabels('')
# If cluster 1 is the same as cluster 2, get the autocorrelation
else:
plt.autocorr(t1,
ax=ax,
color=colors[cl1],
bin_width=bin_width,
limit=limit)
ax.set_ylabel('{}'.format(cl1))
ax.set_xticklabels('')
ax.set_yticklabels('')
return fig
def autocorrs(self,
clusters=None,
bin_width=0.0015,
limit=0.03,
figsize=None):
""" Plots of autocorrelations of clustered spike times.
**Keywords**:
*clusters*: list or iterable
List of clusters to plot
*bin_width*: float
Width of bins in the autocorrelation calculation
*limit*: float
Time limit over which to calculate the autocorrelation
"""
cls = self.clusters.select(clusters)
cl_times = cls.times()
colors = plt.get_colors(max(self.clusters.keys()) + 1, self.cm)
fig, axes = plt.generate_axes(len(cls),
4,
num=3,
figsize=figsize,
sharex=True)
for ax, cl in zip(axes, cl_times):
ax.clear()
tstamps = cl_times[cl]
plt.autocorr(tstamps,
ax=ax,
color=colors[cl],
bin_width=bin_width,
limit=limit)
ax.set_title('Cluster {}'.format(cl))
ax.set_xlabel('Lag (ms)')
fig.tight_layout()
return fig
def plot_stacked_country_cft(self, df, arr_legend, path_out, fname, ncols=2, xlabel='', ylabel='', title=''):
"""
:param df:
:param arr_legend:
:param path_out:
:param fname:
:param xlabel:
:param ylabel:
:param title:
:return:
"""
df = (df
.reset_index()
.fillna(0.0))
ax = (df
.plot
.bar(stacked=True, colormap=plots.get_colors(palette='tableau', cmap=True), linewidth=0, use_index=False))
# Legend
leg = ax.legend(fancybox=None, ncol=ncols, prop={'size': 6})
leg.get_frame().set_linewidth(0.0)
leg.get_frame().set_alpha(0.5)
# Create nice-looking grid for ease of visualization
ax.grid(which='minor', alpha=0.2, linestyle='--')
ax.grid(which='major', alpha=0.5, linestyle='--')
plt.xticks(df.index, df[self.FAO_code])
plt.ylabel(ylabel)
plt.xlabel(xlabel)
plt.title(title)
plt.tight_layout()
plt.savefig(path_out + os.sep + fname, dpi=constants.DPI)
plt.close()
def plot_cnt_decade(inp_fao_df,cnt):
out_dec_df = per_CFT_by_decade(inp_fao_df,cnt)
out_ann_df = per_CFT_annual(inp_fao_df,cnt)
out_dec_df = out_dec_df.set_index('functional_crop_type')
ax = out_dec_df.drop('country_name', axis=1).T.plot(kind='bar',stacked=True,color=plots.get_colors(5),linewidth=0)
plots.simple_axis(ax) # Simple axis, no axis on top and right of plot
# Transparent legend in lower left corner
leg = plt.legend(loc='lower left',fancybox=None)
leg.get_frame().set_linewidth(0.0)
leg.get_frame().set_alpha(0.5)
# Set X and Y axis labels and title
ax.set_title(cnt)
ax.set_xlabel('')
plt.ylim(ymax = 100)
ax.set_ylabel('Percentage of cropland area \noccupied by each crop functional type')
fmt = '%.0f%%' # Format you want the ticks, e.g. '40%'
yticks = mtick.FormatStrFormatter(fmt)
ax.yaxis.set_major_formatter(yticks)
# remove ticks from X axis
plt.tick_params(\
axis='x', # changes apply to the x-axis
which='both', # both major and minor ticks are affected
bottom='off', # ticks along the bottom edge are off
top='off') # ticks along the top edge are off
# Rotate the X axis labels to be horizontal
locs, labels = plt.xticks()
plt.setp(labels, rotation=0)
plt.tight_layout()
plt.savefig(constants.out_dir+os.sep+cnt+'.png', bbox_inches='tight', dpi=600)
plt.close()
