def plot_mean_boxplot_with_pearson(dataset_id):
data = []
pearson = []
for i, technique_id in enumerate(technique_list):
print(Globals.acronyms[technique_id], end=' ', flush=True)
technique_pearson = []
technique_data = []
history = Parser.parse_rectangles(technique_id, dataset_id)
for revision in range(len(history) - 1):
delta_vis = DeltaMetrics.compute_delta_vis(history[revision],
history[revision + 1])
delta_data = DeltaMetrics.compute_delta_data(
history[revision], history[revision + 1])
un_mov = UnavoidableMovement.compute_unavoidable_movement(
history[revision], history[revision + 1])
ratios = (1 - delta_vis) / (1 - delta_data)
diffs = 1 - abs(delta_vis - delta_data)
unavoidable = 1 - (delta_vis - un_mov)
mean = (ratios + diffs + unavoidable) / 3
technique_data.append(mean)
# Compute linear regression statistics
_, _, r_value, _, _ = stats.linregress(delta_data, delta_vis)
technique_pearson.append(r_value if r_value > 0 else 0)
data.append(technique_data)
pearson.append(technique_pearson)
TimeBoxplot.plot_with_pearson(data,
technique_list,
pearson,
title='Mean with Pearson - ' + dataset_id)
def plot_time_boxplot(dataset_id):
data = []
for i, technique_id in enumerate(technique_list):
technique_data = []
history = Parser.parse_rectangles(technique_id, dataset_id)
for revision in range(len(history) - 1):
shneiderman = compute_shneiderman(history[revision],
history[revision + 1])
technique_data.append(shneiderman)
data.append(technique_data)
TimeBoxplot.plot(data, technique_list, title="Shneiderman - " + dataset_id)
def plot_time_boxplot(dataset_id):
data = []
for i, technique_id in enumerate(technique_list):
technique_data = []
history = Parser.parse_rectangles(technique_id, dataset_id)
for revision in range(len(history) - 1):
rpc = relative_position_change_wrapper(history[revision],
history[revision + 1])
technique_data.append(rpc)
data.append(technique_data)
print(Globals.acronyms[technique_id], end=' ', flush=True)
TimeBoxplot.plot(data, technique_list, title="RPC - " + dataset_id)
def pearson_matrix(dataset_ids):
matrix = []
for dataset_id in dataset_ids:
dataset_values = []
for technique_id in technique_list:
# print(Globals.acronyms[technique_id], dataset_id)
history = Parser.parse_rectangles(technique_id, dataset_id)
# Compute all delta_vis and delta_data values for a dataset (1 pair per cell)
all_delta_data = np.array([])
all_delta_vis = np.array([])
for revision in range(len(history) - 1):
delta_data = DeltaMetrics.compute_delta_data(history[revision], history[revision + 1])
all_delta_data = np.append(all_delta_data, delta_data)
delta_vis = DeltaMetrics.compute_delta_vis(history[revision], history[revision + 1])
all_delta_vis = np.append(all_delta_vis, delta_vis)
# Compute linear regression statistics
slope, intercept, r_value, p_value, std_err = stats.linregress(all_delta_data, all_delta_vis)
dataset_values.append(r_value)
print(Globals.acronyms[technique_id], dataset_id, r_value)
matrix.append(dataset_values)
matrix = np.array(matrix).transpose()
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=False,
cell_text=True,
title='Pearson')
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=True,
cell_text=True,
title='Pearson')
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=False,
cell_text=False,
title='Pearson')
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=True,
cell_text=False,
title='Pearson')
def plot_ar_matrix(dataset_ids):
matrix = []
for dataset_id in dataset_ids:
dataset_values = []
for technique_id in technique_list:
history = Parser.parse_rectangles(technique_id, dataset_id)
all_ratios = np.array([])
for revision in range(len(history) - 1):
ratios = compute_aspect_ratios(history[revision])
all_ratios = np.append(all_ratios, ratios.values)
dataset_values.append(all_ratios.mean())
print(Globals.acronyms[technique_id], dataset_id,
all_ratios.mean())
matrix.append(dataset_values)
matrix = np.array(matrix).transpose()
MatrixPlot.plot(matrix,
dataset_ids,
technique_list,
shared_cm=False,
cell_text=True,
title='Aspect ratio')
MatrixPlot.plot(matrix,
dataset_ids,
technique_list,
shared_cm=True,
cell_text=True,
title='Aspect ratio')
MatrixPlot.plot(matrix,
dataset_ids,
technique_list,
shared_cm=False,
cell_text=False,
title='Aspect ratio')
MatrixPlot.plot(matrix,
dataset_ids,
technique_list,
shared_cm=True,
cell_text=False,
title='Aspect ratio')
def plot_matrix(dataset_ids):
matrix = []
for dataset_id in dataset_ids:
dataset_values = []
for technique_id in technique_list:
history = Parser.parse_rectangles(technique_id, dataset_id)
avg = compute_location_drift(history)
dataset_values.append(avg)
print(Globals.acronyms[technique_id], dataset_id, avg)
matrix.append(dataset_values)
matrix = np.array(matrix).transpose()
MatrixPlot.plot(matrix,
dataset_ids,
technique_list,
cell_text=True,
title='Location Drift')
def delta_ratio_boxplots(dataset_id):
data = []
for i, technique_id in enumerate(technique_list):
technique_data = []
history = Parser.parse_rectangles(technique_id, dataset_id)
for revision in range(len(history) - 1):
delta_vis = compute_delta_vis(history[revision], history[revision + 1])
delta_data = compute_delta_data(history[revision], history[revision + 1])
ratios = (1 - delta_vis) / (1 - delta_data)
technique_data.append(ratios)
data.append(technique_data)
TimeBoxplot.plot(data, technique_list,
title="Delta Ratio - " + dataset_id)
TimeBoxplot.plot(data, technique_list,
median_sorted=True,
title="Delta Ratio - " + dataset_id)
def plot_time_boxplot(dataset_id):
data = []
for i, technique_id in enumerate(technique_list):
technique_data = []
history = Parser.parse_rectangles(technique_id, dataset_id)
for revision in range(len(history)):
ratios = compute_aspect_ratios(history[revision]).tolist()
technique_data.append(ratios)
data.append(technique_data)
TimeBoxplot.plot(data,
technique_list,
title="Aspect Ratios - " + dataset_id)
TimeBoxplot.plot(data,
technique_list,
median_sorted=True,
title="Aspect Ratios - " + dataset_id)
def plot_mean_boxplot(
dataset_id,
metrics='VIS'): # Default case was what was used at VIS18 paper
data = []
for i, technique_id in enumerate(technique_list):
print(Globals.acronyms[technique_id], end=' ', flush=True)
technique_data = []
history = Parser.parse_rectangles(technique_id, dataset_id)
for revision in range(len(history) - 1):
if metrics == 'VIS':
delta_vis = DeltaMetrics.compute_delta_vis(
history[revision], history[revision + 1])
delta_data = DeltaMetrics.compute_delta_data(
history[revision], history[revision + 1])
un_mov = UnavoidableMovement.compute_unavoidable_movement(
history[revision], history[revision + 1])
ratios = (1 - delta_vis) / (1 - delta_data)
diffs = 1 - abs(delta_vis - delta_data)
unavoidable = 1 - (delta_vis - un_mov)
mean = (ratios + diffs + unavoidable) / 3
elif metrics == 'SIBGRAPI':
delta_vis = DeltaMetrics.compute_delta_vis(
history[revision], history[revision + 1])
delta_data = DeltaMetrics.compute_delta_data(
history[revision], history[revision + 1])
ratios = (1 - delta_vis) / (1 - delta_data)
shn = ShneidermanWattenberg.compute_shneiderman(
history[revision], history[revision + 1])
mean = (ratios + shn) / 2
technique_data.append(mean)
data.append(technique_data)
TimeBoxplot.plot(data, technique_list, title='Mean - ' + dataset_id)
TimeBoxplot.plot(data,
technique_list,
median_sorted=True,
title='Mean - ' + dataset_id)
def plot(dataset_id):
fig, axs = plt.subplots(math.ceil(len(technique_list) / 5), 5)
fig.suptitle("Centroid Trail - " + dataset_id)
norm = math.sqrt(
1000**2 + 1000**
2) # Assuming we are plotting the treemap in a 1000x1000 pixel frame
for i, technique_id in enumerate(technique_list):
ax = fig.axes[i]
ax.set_title(Globals.acronyms[technique_id])
history = Parser.parse_rectangles(technique_id, dataset_id)
centroids = compute_centroids(history)
lines = []
colors = []
for key, centroid_list in centroids.items():
for i in range(len(centroid_list) - 1):
a = centroid_list[i]
b = centroid_list[i + 1]
lines.append((a, b)) # Add line segment
alpha = math.sqrt((a[0] - b[0])**2 + (a[1] - b[1])**2) / norm
alpha = alpha / math.sqrt(len(centroid_list))
colors.append((0, 0, 0, alpha)) # Set color for line segment
lc = mc.LineCollection(lines, colors=colors, linewidths=1)
ax.add_collection(lc)
ax.set_xlim(0, 1000)
ax.set_ylim(0, 1000)
ax.tick_params(axis='x',
which='both',
bottom=False,
top=False,
labelbottom=False)
ax.tick_params(axis='y',
which='both',
left=False,
right=False,
labelleft=False)
ax.set_aspect('equal', adjustable='box')
fig.savefig(Globals.plot_subdir + dataset_id + '.png', dpi=500)
def delta_ratio_matrix(dataset_ids):
matrix = []
for dataset_id in dataset_ids:
dataset_values = []
for technique_id in technique_list:
history = Parser.parse_rectangles(technique_id, dataset_id)
all_ratios = np.array([])
for revision in range(len(history) - 1):
delta_vis = compute_delta_vis(history[revision], history[revision + 1])
delta_data = compute_delta_data(history[revision], history[revision + 1])
ratio = (1 - delta_vis) / (1 - delta_data)
all_ratios = np.append(all_ratios, ratio.values)
dataset_values.append(all_ratios.mean())
print(Globals.acronyms[technique_id], dataset_id, all_ratios.mean())
matrix.append(dataset_values)
matrix = np.array(matrix).transpose()
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=False,
cell_text=True,
title='Delta ratio')
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=True,
cell_text=True,
title='Delta ratio')
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=False,
cell_text=False,
title='Delta ratio')
MatrixPlot.plot(matrix, dataset_ids, technique_list,
shared_cm=True,
cell_text=False,
title='Delta ratio')
def plot_matrix(dataset_ids):
matrix = []
for dataset_id in dataset_ids:
dataset_values = []
for technique_id in technique_list:
history = Parser.parse_rectangles(technique_id, dataset_id)
all_ratios = np.array([])
for revision in range(len(history) - 1):
distances = compute_shneiderman(history[revision],
history[revision + 1])
all_ratios = np.append(all_ratios, distances.values)
dataset_values.append(all_ratios.mean())
print(Globals.acronyms[technique_id], dataset_id,
all_ratios.mean())
matrix.append(dataset_values)
matrix = np.array(matrix).transpose()
MatrixPlot.plot(matrix,
dataset_ids,
technique_list,
cell_text=True,
title='Shneiderman-Wattenberg')
def scatter(dataset_id):
fig, axs = plt.subplots(int(len(technique_list)/2), 2, sharex=True, sharey=True, figsize=(20, 44))
xlim = 0
for i, technique_id in enumerate(technique_list):
# print(Globals.acronyms[technique_id])
ax = fig.axes[i]
history = Parser.parse_rectangles(technique_id, dataset_id)
# Compute all delta_vis and delta_data values for a dataset (1 pair per cell)
all_delta_data = np.array([])
all_delta_vis = np.array([])
for revision in range(len(history) - 1):
delta_data = DeltaMetrics.compute_delta_data(history[revision], history[revision + 1])
all_delta_data = np.append(all_delta_data, delta_data)
delta_vis = DeltaMetrics.compute_delta_vis(history[revision], history[revision + 1])
all_delta_vis = np.append(all_delta_vis, delta_vis)
# Compute linear regression and draw regression line
slope, intercept, r_value, p_value, std_err = stats.linregress(all_delta_data, all_delta_vis)
if xlim == 0:
xlim = np.percentile(all_delta_data, 99.99) # Remove outliers
# If there are too many points to handle, first we draw all of them in black (alpha),
# then subsample the space, perform kde, and draw the colored subsample
# over the original points
sample_size = 10000
if len(all_delta_data) > sample_size:
ax.scatter(all_delta_data, all_delta_vis, color='k', s=1, alpha=.25)
# matrix = df[['delta_data', 'delta_vis']].sample(sample_size).T.as_matrix()
indices = np.random.choice(len(all_delta_vis), sample_size)
matrix = np.vstack([all_delta_data[indices], all_delta_vis[indices]])
dens = stats.gaussian_kde(matrix)
dens_pt = dens(matrix)
colours = make_colors(dens_pt, 'inferno')
ax.scatter(matrix[0], matrix[1], color=colours, s=3, alpha=.05)
else:
matrix = np.vstack([all_delta_data, all_delta_vis])
dens = stats.gaussian_kde(matrix)
dens_pt = dens(matrix)
colours = make_colors(dens_pt, 'inferno')
ax.scatter(matrix[0], matrix[1], color=colours, s=3, alpha=.25)
line = np.poly1d([slope, intercept])(all_delta_data)
ax.plot(all_delta_data, line, 'r-', lw=.5)
print(Globals.acronyms[technique_id], r_value)
title = Globals.acronyms[technique_id]
title += r" $\alpha = $" + "{0:.2f}".format(intercept)
title += r" $\beta = $" + "{0:.2f}".format(slope)
title += r" $r = $" + "{0:.3f}".format(r_value)
title += r" $s_e = $" + "{0:.3f}".format(std_err)
ax.set_title(title)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.tick_params(axis='x', which='both', top='off')
ax.tick_params(axis='y', which='both', right='off')
ax.set_xlim(xmin=0, xmax=xlim)
ax.set_ylim(ymin=0)
fig.savefig(Globals.plot_subdir + 'scatter-' + dataset_id + '.png', dpi=500)
# plt.show()
return
