def scatter_density(x,
y,
res=100,
cmap=matplotlib.cm.hot_r,
size=None,
return_cbar=False,
**kwargs):
"""
Create a scatter plot with density of the data-points at each point on the
x,y grid coded by the color in the colormap (hot per default)
**kwargs are passed to matplotlib's matshow
"""
x = np.copy(x)
y = np.copy(y)
max_x = np.nanmax(x)
max_y = np.nanmax(y)
min_x = np.nanmin(x)
min_y = np.nanmin(y)
x = ozu.rescale(x).ravel() * (res - 1)
y = ozu.rescale(y).ravel() * (res - 1)
data_arr = np.zeros((res, res))
for this_x, this_y in zip(x, y):
# If one of them is a nan, move on:
if np.isnan(this_x) or np.isnan(this_y):
pass
else:
data_arr[int(np.floor(this_x)), int(np.floor(this_y))] += 1
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
imax = ax.matshow(np.log10(np.flipud(data_arr.T)), cmap=cmap, **kwargs)
cbar = fig.colorbar(imax)
ax.set_xticks([0] + [i * res / 5.0 for i in range(5)])
ax.set_yticks([0] + [i * res / 5.0 for i in range(5)])
ax.set_xticklabels(
[0] +
['%0.2f' % (i * ((max_x - min_x) / 5.0) + min_x) for i in range(5)])
ax.set_yticklabels([0] + [
'%0.2f' % (i * ((max_y - min_y) / 5.0) + min_y)
for i in range(5, 0, -1)
])
if size is not None:
fig.set_size_inches(size)
if return_cbar:
return fig, cbar
else:
return fig
def test_rescale():
"""
Test rescaling of data into [0,1]
"""
data = np.random.randn(20) * 100
rs = ozu.rescale(data)
npt.assert_equal(np.max(rs), 1)
npt.assert_equal(np.min(rs), 0)
# Test for conditions in which the minimum is >0:
data = (np.random.rand(20) + 10) * 100
rs = ozu.rescale(data)
npt.assert_equal(np.max(rs), 1)
npt.assert_equal(np.min(rs), 0)
def test_rescale():
"""
Test rescaling of data into [0,1]
"""
data = np.random.randn(20) * 100
rs = ozu.rescale(data)
npt.assert_equal(np.max(rs),1)
npt.assert_equal(np.min(rs),0)
# Test for conditions in which the minimum is >0:
data = (np.random.rand(20) + 10) * 100
rs = ozu.rescale(data)
npt.assert_equal(np.max(rs),1)
npt.assert_equal(np.min(rs),0)
def scatter_density(x,y, res=100, cmap=matplotlib.cm.hot_r, size=None,
return_cbar=False, **kwargs):
"""
Create a scatter plot with density of the data-points at each point on the
x,y grid coded by the color in the colormap (hot per default)
**kwargs are passed to matplotlib's matshow
"""
x = np.copy(x)
y = np.copy(y)
max_x = np.nanmax(x)
max_y = np.nanmax(y)
min_x = np.nanmin(x)
min_y = np.nanmin(y)
x = ozu.rescale(x).ravel() * (res - 1)
y = ozu.rescale(y).ravel() * (res - 1)
data_arr = np.zeros((res, res))
for this_x,this_y in zip(x,y):
# If one of them is a nan, move on:
if np.isnan(this_x) or np.isnan(this_y):
pass
else:
data_arr[int(np.floor(this_x)), int(np.floor(this_y))] += 1
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
imax = ax.matshow(np.log10(np.flipud(data_arr.T)), cmap=cmap, **kwargs)
cbar = fig.colorbar(imax)
ax.set_xticks([0] + [i * res/5.0 for i in range(5)])
ax.set_yticks([0] + [i * res/5.0 for i in range(5)])
ax.set_xticklabels([0] + ['%0.2f'%(i * ((max_x - min_x)/5.0) + min_x)
for i in range(5)])
ax.set_yticklabels([0] + ['%0.2f'%(i * ((max_y - min_y)/5.0) + min_y)
for i in range(5,0,-1)])
if size is not None:
fig.set_size_inches(size)
if return_cbar:
return fig, cbar
else:
return fig