def plotBar(pos, cases, dt, color, title):
"""
This plots based base on the time series data
:param pos:
:param cases:
:param dt:
:param color:
:param title:
:return:
"""
labels = [val[1] for val in cases]
labels2 = [
"Test"
] + labels # this is to overcome a bug in the bar function
x = np.arange(len(labels)) # the label locations
axs[pos].bar(x, [val[0] for val in cases],
color=color,
edgecolor='k')
axs[pos].set_title("{} on {}".format(title, str(dt)), fontsize=15)
axs[pos].xaxis.set_major_locator(plt.FixedLocator(x))
axs[pos].xaxis.set_major_formatter(plt.FixedFormatter(labels))
#axs[pos].set_xticklabels(labels2)
axs[pos].grid(axis="y")
plt.setp(axs[pos].xaxis.get_majorticklabels(),
rotation=35,
fontsize=12)
def graph_flow_generator(total):
ticket_dict = {}
ticket_dict["Information"] = total[0]
ticket_dict["Mineure"] = total[1]
ticket_dict["Majeure"] = total[2]
ticket_dict["Bloquante"] = total[3]
ticket_dict["Autre"] = total[4]
fig = plt.figure()
ax = fig.add_subplot(111)
frequencies = ticket_dict.values()
names = ticket_dict.keys()
x_coordinates = np.arange(len(ticket_dict))
ax.bar(x_coordinates, frequencies, align='center')
ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.xaxis.set_major_formatter(plt.FixedFormatter(names))
plt.savefig("flow" + contract_id + ".svg")
# plt.show()
img_insert("Repartition des demandes", "flow" + contract_id + ".svg",
"Repartition des demandes")
def bar_plot(d, bars, outdir='.'):
N = len(bars)
ind = numpy.arange(N)
width = 0.35
plt.title("LCA of reads for contigs with alignment purity < 0.95", size='x-small')
# add bars
sum_values = sum(d.values())
plt.bar(ind, [d[k] * 100 / sum_values for k in bars], width, color='red')
# Final bar is not really visible, increase xlim
xmin, xmax = plt.xlim()
plt.xlim(xmin=xmin-1)
plt.xlim(xmax=xmax+1)
# axis setup
labels = []
for k in bars:
labels.append('%s\n%d' % (k, d[k]))
plt.xticks(ind + width / 2., bars, size='xx-small', rotation=17)
tick_range = numpy.arange(0, 110, 10)
plt.yticks(tick_range, size='xx-small')
formatter = plt.FixedFormatter([str(x) for x in tick_range])
plt.gca().yaxis.set_major_formatter(formatter)
plt.gca().yaxis.grid(which='major')
plt.savefig(outdir + "/plots/barplot.png")
plt.close()
def generate_histrogram_strings(buckets, title, x_label, y_label, output):
fig = plt.figure()
ax = fig.add_subplot(111)
sorted_keys = sorted(buckets.keys(),
key=lambda item: int(''.join(x for x in item
if x.isdigit())))
indexes = np.arange(0, len(sorted_keys)) # the x locations for the groups
width = 0.1 # the width of the bars
ax.set_yscale('log')
ax.grid(zorder=0, axis="y")
bars1 = ax.bar(indexes, [buckets.get(sorted_keys[i], 0) for i in indexes],
align='center',
color='blue')
# axes and labels
ax.set_ylim(0.5, max(buckets.values()) + 1)
ax.set_xlabel(x_label)
ax.set_ylabel(y_label)
ax.set_title(title)
ax.xaxis.set_major_locator(plt.FixedLocator(indexes))
ax.xaxis.set_major_formatter(plt.FixedFormatter(sorted_keys))
labels = ax.get_xticklabels()
for tick in labels:
tick.set_rotation(90)
plt.tight_layout()
fig.savefig(output)
def plot_bar_graph_within(names, observered_data, result_data, title, fig,
sub_plot_index):
"""
:param names: labels of x-axis
:param observered_data: as variable name
:param result_data: as variable name
:param title: the title of the subplot
:param fig: the fig object
:param sub_plot_index: the index number is in order,
such 221 means the generate a graph has 2 by 2 figure, which could have 4 subplots in it
the 1 means the first subplot in the figure.
:return: the fig object
"""
# define the subplot
ax = fig.add_subplot(sub_plot_index)
x = np.arange(names.__len__()) * 2
# set the value
ax.bar(x, observered_data, align='center', color='red')
ax.bar(x + 1, result_data, align='center', color='blue')
# create the patch
red_patch = mpatches.Patch(color='red', label='Observed Value')
blue_patch = mpatches.Patch(color='blue', label='Simulated Value')
# Set the label x-axis
ax.xaxis.set_major_locator(plt.FixedLocator(x + 0.5))
ax.xaxis.set_major_formatter(plt.FixedFormatter(names))
ax.set_title(title)
plt.legend(handles=[red_patch, blue_patch])
return fig
def save_ability_power_statistics(statistics):
fig = plt.figure()
ax = fig.add_subplot(111)
x = np.arange(len(statistics))
plt.bar(x, [s[1] for s in statistics], width=0.8, align='center')
ax.set_xlim(-0.5, len(statistics) + 0.5)
ax.set_ylim(0, statistics[0][1])
locator = plt.IndexLocator(1, 0.4)
formatter = plt.FixedFormatter([s[0] for s in statistics])
ax.xaxis.set_major_locator(locator)
ax.xaxis.set_major_formatter(formatter)
plt.setp(plt.getp(ax, 'xticklabels'),
rotation=45,
fontsize=8,
horizontalalignment='right')
ax.set_title('Wins per ability')
plt.tight_layout()
plt.savefig('/tmp/wins.png')
def plot_bar_from_counter(counter, title, xlabel, ylabel, ax=None):
""""
This function creates a bar plot from a counter.
:param counter: This is a counter object, a dictionary with the item as the key
and the frequency as the value
:param ax: an axis of matplotlib
:return: the axis wit the object in it
"""
if ax is None:
fig = plt.figure()
ax = fig.add_subplot(111)
frequencies = counter.values()
names = list(counter.keys())
x_coordinates = np.arange(len(counter))
ax.bar(x_coordinates, frequencies, align='center')
ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.xaxis.set_major_formatter(plt.FixedFormatter(names))
ax.xaxis.set_label_text(xlabel)
ax.yaxis.set_label_text(ylabel)
ax.set_title(title)
plt.autoscale(enable=True, axis='x')
DefaultSize = fig.get_size_inches()
fig.set_size_inches((DefaultSize[0] * 5, DefaultSize[1] * 2.5))
mkdir_p(IMG_URL)
plt.savefig(IMG_URL + 'GIST Data Histogram.png')
plt.close()
im = crop_image(IMG_URL + 'GIST Data Histogram.png')
return ax
def save_ability_wins_distribution(statistics, ability_wins):
fig = plt.figure()
ax = fig.add_subplot(111)
keys, wins = list(zip(*statistics)) # pylint: disable=W0612
data = [ability_wins[key] for key in keys]
ax.boxplot(data) #, positions=[i for i in xrange(len(keys))])
ax.set_xlim(0.5, len(statistics) + 0.5)
ax.set_ylim(0, TEST_BATTLES_NUMBER * len(HERO_LEVELS))
locator = plt.IndexLocator(1, 0.5)
formatter = plt.FixedFormatter([s[0] for s in statistics])
ax.xaxis.set_major_locator(locator)
ax.xaxis.set_major_formatter(formatter)
plt.setp(plt.getp(ax, 'xticklabels'),
rotation=45,
fontsize=8,
horizontalalignment='right')
ax.set_title('wins destribution')
plt.tight_layout()
plt.savefig('/tmp/wins_destribution.png')
def plot_bar_from_counter(counter, ax=None):
""""
This function creates a bar plot from a counter.
:param counter: This is a counter object, a dictionary with the item as the key
and the frequency as the value
:param ax: an axis of matplotlib
:return: the axis wit the object in it
"""
if ax is None:
fig = plt.figure(figsize=(7, 5))
ax = fig.add_subplot(111)
frequencies = [f[1] for f in counter]
names = [f[0] for f in counter]
x_coordinates = np.arange(len(counter))
ax.barh(x_coordinates, frequencies, align='center')
ax.yaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.yaxis.set_major_formatter(plt.FixedFormatter(names))
ax.yaxis.set_tick_params(labelsize=10)
ax.set_ylim(-1, len(counter) + 1)
ax.set_xlim(0, max(frequencies) + 300)
for i, v in enumerate(frequencies):
ax.text(v + 20,
i - 0.25,
str(v),
color='black',
fontweight='normal',
fontsize=10)
fig.subplots_adjust(left=0.4)
return ax
def ohlc_plot(ax, df, open_='open', high_='high', low_='low', close_='close', t_='trading_day', width_=0.7, n_=10):
assert isinstance(df, pd.DataFrame)
assert all([x in df.columns for x in (open_, high_, low_, close_)])
n = len(df)
width = width_
offset = width / 2.0
for i in range(n):
op = df[open_].values[i]
hi = df[high_].values[i]
lo = df[low_].values[i]
cl = df[close_].values[i]
x = i
y = min(op, cl)
height = abs(cl - op)
is_raise = op < cl
clr = 'red' if is_raise else 'green'
fill = True # True if is_raise else False
rect = Rectangle((x - offset, y), width, height, facecolor=clr, edgecolor=clr, linewidth=1, fill=fill)
vline = Line2D(xdata=(x, x), ydata=(lo, hi), linewidth=1, color=clr)
ax.add_patch(rect)
ax.add_line(vline)
ax.autoscale_view()
x_label_idx = n_ * np.array(range(len(df) // n_ + 1))
x_labels = df[t_].iloc[x_label_idx].values
ax.xaxis.set_major_locator(plt.FixedLocator(x_label_idx))
ax.xaxis.set_major_formatter(plt.FixedFormatter(x_labels))
def plot_bar_from_counter(counter, ax=None):
""""
This function creates a bar plot from a counter.
:param counter: This is a counter object, a dictionary with the item as the key
and the frequency as the value
:param ax: an axis of matplotlib
:return: the axis wit the object in it
"""
if ax is None:
fig = plt.figure()
ax = fig.add_subplot(111)
frequencies = counter.values()
names = counter.keys()
x_coordinates = np.arange(len(counter))
ax.bar(x_coordinates, frequencies, align='center')
for idx in xrange(len(frequencies)):
x = x_coordinates[idx]
y = frequencies[idx]
ax.text(x + 3, y + 3, str(v), color='red', fontweight='bold')
ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.xaxis.set_major_formatter(plt.FixedFormatter(names))
ax.set_ylim(0, np.max(frequencies) * 1.1)
return ax
def climate_plot():
df1 = climate_data()
df1.name = 'Total GHG'
df1.index = pd.to_datetime(df1.index, format='%Y') # 將index改為年datetime object
df1.index = df1.index.to_period('A').to_timestamp('A') # 將datetime start of the year change to end of the year
df_temp = df_temperature
df_temp.index = pd.to_datetime(df_temp.index, format='%Y-%m-%d') # 將index改為年datetime object
df2 = df_temp
df2 = df2.resample('A').mean() # 重採樣為以年為頻率，方法為平均
df_temp = df_temp.resample('Q').mean()
df_ghg_temp = normalize(pd.concat([df2, df1], axis=1, join='inner')).loc['1990':'2010'] # 1990 年 - 2010 年
fig, axes = plt.subplots(nrows=2, ncols=2)
# 子图 1（线形图 kind = 'line'）：绘制 1990 年 - 2010 年间，「全球历年温室气体排放总量」与「历年陆地平均温度」及「历年陆地-海洋平均温度」三者之间的线形图。
#plt.subplot(221)
ax1 = df_ghg_temp.plot(kind='line', ax=axes[0, 0], figsize=(16, 9))
ax1.set_xlabel('Years')
ax1.set_ylabel('Values')
#ax1.legend(loc='best')
ax1.autoscale(axis='x', tight=True)
ax1.grid(True)
# 子图 2（柱状图 kind = 'bar'）：绘制 1990 年 - 2010 年间，「全球历年温室气体排放总量」与「历年陆地平均温度」及「历年陆地-海洋平均温度」三者之间的柱状图。
#plt.subplot(222)
ax2 = df_ghg_temp.plot(kind='bar', ax=axes[0, 1], figsize=(16, 9))
ax2.xaxis.set_major_formatter(plt.FixedFormatter(df_ghg_temp.index.to_series().dt.strftime("%Y")))
ax2.set_xlabel('Years')
ax2.set_ylabel('Values')
#ax2.legend(loc='best')
ax2.autoscale(axis='x', tight=True)
ax2.grid(True)
# 子图 3（面积图 kind = 'area'）：绘制有气象数据的年份，「各季度地面平均温度」与「各季度地面-海洋平均温度」面积图。
#plt.subplot(223)
ax3 = df_temp.plot(kind='area', ax=axes[1, 0], figsize=(16, 9))
ax3.set_xlabel('Quarters')
ax3.set_ylabel('Temperature')
#ax3.legend(loc='best')
ax3.autoscale(axis='x', tight=True)
ax3.grid(True)
# 子图 4（核密度估计图 kind = 'kde'）：绘制有气象数据年份，「各季度地面平均温度」与「各季度地面-海洋平均温度」对应的核密度估计图。
#plt.subplot(224)
ax4 = df_temp.plot(kind='kde', ax=axes[1, 1], figsize=(16, 9)) # set figsize 避免擠在一起
ax4.set_xlabel('Values')
ax4.set_ylabel('Values')
ax4.autoscale(axis='x', tight=True)
ax4.grid(True)
# Adjust the subplot layout, because the logit one may take more space
# than usual, due to y-tick labels like "1 - 10^{-3}"
plt.subplots_adjust(top=0.92, bottom=0.08, left=0.10, right=0.95, hspace=0.25,
wspace=0.35)
plt.show()
return fig
def palplot(k, cmap="viridis"):
pal = sns.color_palette(palette=cmap, n_colors=k)
fig, ax = plt.subplots(1, 1)
pal = np.array(pal)
pal = pal.reshape((k, 1, 3))
ax.imshow(pal)
ax.xaxis.set_major_locator(plt.NullLocator())
ax.yaxis.set_major_formatter(plt.FixedFormatter(np.arange(k, dtype=int)))
ax.yaxis.set_major_locator(plt.FixedLocator(np.arange(k)))
return ax
def plot_bar_from_counter(counter):
ax = fig.add_subplot(1, 1, 1)
x_coordinates = np.arange(len(counter))
ax.bar(x_coordinates, counter.values(), align='center')
ax.xaxis.set_tick_params(rotation=90)
ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.xaxis.set_major_formatter(plt.FixedFormatter(list(counter.keys())))
return ax
def test_twin_axis_locaters_formatters():
vals = np.linspace(0, 1, num=5, endpoint=True)
locs = np.sin(np.pi * vals / 2.0)
majl = plt.FixedLocator(locs)
minl = plt.FixedLocator([0.1, 0.2, 0.3])
fig = plt.figure()
ax1 = fig.add_subplot(1, 1, 1)
ax1.plot([0.1, 100], [0, 1])
ax1.yaxis.set_major_locator(majl)
ax1.yaxis.set_minor_locator(minl)
ax1.yaxis.set_major_formatter(plt.FormatStrFormatter('%08.2lf'))
ax1.yaxis.set_minor_formatter(plt.FixedFormatter(['tricks', 'mind', 'jedi']))
ax1.xaxis.set_major_locator(plt.LinearLocator())
ax1.xaxis.set_minor_locator(plt.FixedLocator([15, 35, 55, 75]))
ax1.xaxis.set_major_formatter(plt.FormatStrFormatter('%05.2lf'))
ax1.xaxis.set_minor_formatter(plt.FixedFormatter(['c', '3', 'p', 'o']))
ax2 = ax1.twiny()
ax3 = ax1.twinx()
def create_histogram(self): figure = plt.figure() ax = figure.add_subplot(111) frequencies = self.ticket_dict.values() names = self.ticket_dict.keys() x_coordinates = np.arange(len(self.ticket_dict)) ax.bar(x_coordinates , frequencies , align='center') ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates)) ax.xaxis.set_major_formatter(plt.FixedFormatter(names))
def test_learning_rates(train_set_x, train_set_y, test_set_x, test_set_y):
"""
Plot the learning curve and the accuracy varying the learning rate
Arguments:
X_train -- training set represented by a numpy array of shape (number of features, m_train)
Y_train -- training labels represented by a numpy array (vector) of shape (1, m_train)
X_test -- test set represented by a numpy array of shape (number of features, m_test)
Y_test -- test labels represented by a numpy array (vector) of shape (1, m_test)
"""
learning_rates = [0.01, 0.001, 0.0001, 0.05, 0.005]
train_accuracy = []
test_accuracy = []
models = {}
for i in learning_rates:
print("learning rate is: " + str(i))
models[str(i)] = logistic_regression(train_set_x,
train_set_y,
test_set_x,
test_set_y,
epochs=1500,
learning_rate=i)
train_accuracy.append(models[str(i)]["train_accuracy"])
test_accuracy.append(models[str(i)]["test_accuracy"])
print('\n' +
"-------------------------------------------------------" + '\n')
for i in learning_rates:
plt.plot(np.squeeze(models[str(i)]["costs"]),
label=str(models[str(i)]["learning_rate"]))
#plot learning curves
plt.title('Learning curve per Learning rate')
plt.ylabel('cost')
plt.xlabel('iterations (hundreds)')
legend = plt.legend(loc='upper center', shadow=True)
frame = legend.get_frame()
frame.set_facecolor('0.90')
plt.show()
#plot accuracy bar graph
df = pd.DataFrame({
"Train accuracy": train_accuracy,
"Test accuracy": test_accuracy
})
ax = df.plot.bar(color=["SkyBlue", "LightGreen"],
rot=0,
title="Accuracy per learning rate")
ax.set_xlabel("Learning rate")
ax.set_ylabel("Accuracy")
ax.xaxis.set_major_formatter(plt.FixedFormatter(learning_rates))
plt.show()
def plot_daily_data(df, state, type_):
fig = sns.barplot(df.index, df[state])
fig.set_title(f'{type_} Cases ({state})')
fig.set_ylabel('Number of Cases')
fig.set_xticklabels(fig.get_xticklabels(),
rotation=45,
horizontalalignment='right',
fontweight='light')
plt.gca().xaxis.set_major_formatter(
plt.FixedFormatter(df.index.to_series().dt.strftime('%d %b %Y')))
def plot_color_labels(meta, ax):
sizes = meta.groupby(["merge_class"], sort=False).size()
uni_class = sizes.index.unique()
counts = sizes.values
count_map = dict(zip(uni_class, counts))
names = []
colors = []
for key, val in count_map.items():
names.append(f"{key} ({count_map[key]})")
colors.append(CLASS_COLOR_DICT[key])
colors = colors[::-1] # reverse because of signal flow sorting
names = names[::-1]
palplot(len(colors), colors, ax=ax)
ax.yaxis.set_major_formatter(plt.FixedFormatter(names))
def obspred_plot(paramsplot, fityrs, startage, endage,
trans = 'none', trans_time_coords = False, trans_y_coords = False):
obsframe = paramsplot['test'].rx2('obs')
ptype = paramsplot['ptype']
obsage = obsframe.rx2('age')
predage = base.seq(startage, endage, by = 5)
xlinlab = r'$\mathrm{log}(t)$'
ratelinlab = r'$\mathrm{log}(r(t))$'
survlinlab = r'$\mathrm{log}[-\mathrm{log}[S(t)]]$'
plotlabs = {'none': {'x': '$t$', 'rate': '$r(t)$', 'surv': '$S(t)$'},
'gomp_lin': {'x': '$t$', 'rate': ratelinlab, 'surv': survlinlab},
'weib_lin': {'x': xlinlab,'rate': ratelinlab, 'surv': survlinlab}}
plt.close()
fig = plt.figure()
ax = fig.add_subplot(111)
def trans_y_ticks(x, p):
ticks_round = np.round(predcols['none'][ptype], 6)
return str(ticks_round[p]).replace('.', ',')
for yr in fityrs:
obs = obsframe.rx2(str(yr))
pred = stats.predict(paramsplot['test'].rx2('sourcefit').rx2(str(yr)),
newdata = ro.r['list'](age = predage))
obscols = transdict(np.array(obsage), np.array(obs))
predcols = transdict(np.array(predage), np.array(pred))
obsplot = ax.plot(obscols[trans]['x'], obscols[trans][ptype],
'o', label = str(yr))
curcolor = obsplot[0].get_color()
predplot = ax.plot(predcols[trans]['x'], predcols[trans][ptype],
color = curcolor)
plt.legend(loc = 2, framealpha = 0.5)
ax.set_title('Observerad vs förutsedd ' + paramsplot['plottitle'])
if (trans_time_coords):
ax.xaxis.set_major_locator(plt.FixedLocator(predcols[trans]['x']))
ax.xaxis.set_major_formatter(plt.FixedFormatter(predcols['none']['x']))
ax.set_xlabel(plotlabs['none']['x'])
plt.xticks(rotation = 90)
else:
ax.set_xlabel(plotlabs[trans]['x'])
if (trans_y_coords):
ax.yaxis.set_major_locator(plt.FixedLocator(predcols[trans][ptype]))
ax.yaxis.set_major_formatter(plt.FuncFormatter(trans_y_ticks))
ax.set_ylabel(plotlabs['none'][ptype])
else:
ax.set_ylabel(plotlabs[trans][ptype])
def plot_bar_from_counter(counter, ax=None):
if ax is None:
fig = plt.figure()
ax = fig.add_subplot(111)
frequencies = counter.values()
names = counter.keys()
x_coordinates = np.arange(len(counter))
ax.bar(x_coordinates, frequencies, align='center')
ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.xaxis.set_major_formatter(plt.FixedFormatter(names))
return ax
def configure(self):
# Set various axis parameters
self.axis.set_title("Performance of the rat", fontsize=22, pad=10)
self.axis.set_xlabel("Trials", fontsize=16, labelpad=20)
self.axis.set_ylabel("Path length (m)", fontsize=16, labelpad=10)
self.axis.set_ylim(0, TRIAL_TIMEOUT * SWIMING_SPEED * 1.5)
self.axis.tick_params(length=0, labelsize=12)
# Add an horizontal grid to the plot
self.axis.yaxis.grid()
# Group the labels by day using X labels
self.axis.xaxis.set_major_locator(
plt.FixedLocator([2.5 + (i * 4) for i in range(0, 10)]))
self.axis.xaxis.set_major_formatter(
plt.FixedFormatter(["Day " + str(i) for i in range(1, 10)]))
def plot_bar_graph(dictionary, ax=None):
if ax is None:
fig = plt.figure()
ax = fig.add_subplot(111)
counts = dictionary.values()
directories = dictionary.keys()
plt.xlabel('Directories')
plt.ylabel('Counts')
x_coordinates = np.arange(len(dictionary))
ax.bar(x_coordinates, counts, align='center')
ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.xaxis.set_major_formatter(plt.FixedFormatter(directories))
return ax
def plot_class_colormap():
from src.visualization import palplot
from src.data import load_metagraph
mg = load_metagraph("G")
uni_class, counts = np.unique(mg["merge_class"], return_counts=True)
count_map = dict(zip(uni_class, counts))
names = []
colors = []
for key, val in CLASS_COLOR_DICT.items():
if key in uni_class:
names.append(f"{key} ({count_map[key]})")
colors.append(val)
fig, ax = plt.subplots(1, 1, figsize=(3, 15))
palplot(len(colors), colors, ax=ax)
ax.yaxis.set_major_formatter(plt.FixedFormatter(names))
plt.savefig("./maggot_models/notebooks/outs/current_cmap.png",
dpi=300,
bbox_inches="tight")
def plot_hist_days(df, filename=None, col='latitude', figsize=(15,10), timestamp_col='timestamp'):
'''
plot distribution of the data collected on each day
'''
hist_df = df.set_index(pd.DatetimeIndex(df[timestamp_col]), drop=False, inplace=False)
hist_df = hist_df[[col]].groupby(pd.Grouper(freq='D')).count()
sns.set()
fig, ax = plt.subplots(figsize=figsize)
plot = sns.barplot(x=hist_df.index, y=hist_df[col], ax=ax);
ax.set_xlabel('Timestamp');
ax.set_ylabel('Count');
ax.xaxis_date();
ax.xaxis.set_major_formatter(plt.FixedFormatter(hist_df.index.to_series().dt.strftime("%d %b")));
for tick in ax.get_xticklabels():
tick.set_rotation(65);
if filename is not None:
fig = plot.get_figure()
fig.savefig(filename)
def save_ability_mathces_statistics(statistics, matches): # pylint: disable=R0914
fig = plt.figure()
ax = fig.add_subplot(111)
keys, wins = list(zip(*statistics)) # pylint: disable=W0612
index = dict((key, i) for i, key in enumerate(keys))
data = []
for (x, y), (w_1, w_2) in list(matches.items()):
data.append((index[x], index[y], 1000 * w_1 / float(w_1 + w_2)))
data.append((index[y], index[x], 1000 * w_2 / float(w_1 + w_2)))
x, y, powers = list(zip(*data))
ax.scatter(x, y, s=powers, marker='o', c=powers)
ax.set_xlim(-0.5, len(statistics) + 0.5)
ax.set_ylim(-0.5, len(statistics) + 0.5)
locator = plt.IndexLocator(1, 0)
formatter = plt.FixedFormatter([s[0] for s in statistics])
ax.xaxis.set_major_locator(locator)
ax.xaxis.set_major_formatter(formatter)
plt.setp(plt.getp(ax, 'xticklabels'),
rotation=45,
fontsize=8,
horizontalalignment='right')
ax.yaxis.set_major_locator(locator)
ax.yaxis.set_major_formatter(formatter)
plt.setp(plt.getp(ax, 'yticklabels'), fontsize=8)
ax.set_title('matches results')
plt.tight_layout()
plt.savefig('/tmp/matches.png')
def test_epochs(train_set_x, train_set_y, test_set_x, test_set_y):
"""
Plot the accuracy varying the number of epochs
Arguments:
X_train -- training set represented by a numpy array of shape (number of features, m_train)
Y_train -- training labels represented by a numpy array (vector) of shape (1, m_train)
X_test -- test set represented by a numpy array of shape (number of features, m_test)
Y_test -- test labels represented by a numpy array (vector) of shape (1, m_test)
"""
epochs = [100, 500, 1000, 2000, 5000, 10000]
models = {}
train_accuracy = []
test_accuracy = []
for i in epochs:
print("No of epochs: " + str(i))
models[str(i)] = logistic_regression(train_set_x,
train_set_y,
test_set_x,
test_set_y,
epochs=i)
train_accuracy.append(models[str(i)]["train_accuracy"])
test_accuracy.append(models[str(i)]["test_accuracy"])
print('\n' +
"-------------------------------------------------------" + '\n')
#plot accuracy bar graph
df = pd.DataFrame({
"Train accuracy": train_accuracy,
"Test accuracy": test_accuracy
})
ax = df.plot.bar(color=["SkyBlue", "LightGreen"],
rot=0,
title="Accuracy per number of epochs")
ax.set_xlabel("Number of epochs")
ax.set_ylabel("Accuracy")
ax.xaxis.set_major_formatter(plt.FixedFormatter(epochs))
plt.show()
def plot_distrib(train_distrib, val_distrib, test_distrib, info):
fig, ax = plt.subplots()
width = 0.25
x = np.arange(len(info))
ax.bar(x - width, train_distrib, width, label="Trainig")
ax.bar(x, val_distrib, width, label="Validation")
ax.bar(x + width, test_distrib, width, label="Test")
#setting custom labels: put symbols (user input) instead of numbers on the x avis
ax.xaxis.set_major_locator(plt.FixedLocator(x))
ax.xaxis.set_major_formatter(plt.FixedFormatter(info))
#plot configuration
plt.title("Distribution of Instances")
plt.xlabel("Symbol")
plt.ylabel("Frequency")
ax.legend()
ax.grid(True, axis='y', alpha=0.35)
plt.show()
def plot_case_histograms(df, top_n_provinces=6, suptitle=''):
resampled = df.T.iloc[:, :top_n_provinces].resample('d').sum()
print(resampled.tail())
# df.columns = [dt.strftime('%m/%d') for dt in df.columns]
axes = resampled.plot.bar(rot=90,
subplots=True,
sharey=True,
figsize=(6, 8))
for ax in axes:
ax.legend(loc=2)
ax.set(title='')
ax.xaxis.set_major_formatter(
plt.FixedFormatter(resampled.index.to_series().dt.strftime("%d %b")))
fig = plt.gcf()
fig.suptitle(suptitle, y=1)
# fig.autofmt_xdate()
fig.set_tight_layout(True)
def plot_bar_from_counter(counter, ax=None):
""""
This function creates a bar plot from a counter.
:param counter: This is a counter object, a dictionary with the item as the key
and the frequency as the value
:param ax: an axis of matplotlib
:return: the axis wit the object in it
"""
if ax is None:
fig = plt.figure()
ax = fig.add_subplot(111)
frequencies = counter.values()
names = counter.keys()
x_coordinates = np.arange(len(counter))
ax.bar(x_coordinates, frequencies, align='center')
ax.xaxis.set_major_locator(plt.FixedLocator(x_coordinates))
ax.xaxis.set_major_formatter(plt.FixedFormatter(names))
return ax