def draw_histo(df, col_name, x_start, bin_width, n_bins, title, output_dir,
output_filename):
"""
Create a histogram from a pandas dataframe column
Args:
df: a Pandas dataframe
col_name: a string giving the column from the dataframe you'd like to plot
x_start: the leftmost x-value in your plot
bin_width: the desired width of your histogram bins
n_bins: the desired number of histogram bins
title: a string giving the title to be printed on your plot
output_dir: the directory to write your output images to
output_filename: the name you'd like to give the png output image
"""
plt = df[col_name].plot.hist(
bins=[bin_width * (x + 0.5) + x_start for x in np.arange(0, n_bins)],
edgecolor='black',
color='#88d498')
plt.set_title(title)
print(col_name + ' max: ' + str(df[col_name].max()) + ', min: ' +
str(df[col_name].min()))
plt.ticklabel_format(style='sci', axis='x', scilimits=(0, 3))
plt.get_figure().savefig(os.path.join(output_dir, output_filename))
plt.get_figure().clf()
def getPlot(self, params):
param = str(params['sort'])
df = self.getData(params)
x = df[[param]]
y = df['week']
plt = x.plot(y)
fig = plt.get_figure()
return fig
def plotfigTimeSeriesThreetoOne(data):
plt = data.plot(lw=2, title="wind predict before and after MOS ")
plt.set_xlabel("time series")
plt.set_ylabel("wind speed (m/s) ")
fig = plt.get_figure()
fig.savefig('./plot/' + PostObject + '/TimeSeriesThreetoOne.png')
fig.clf()
def sizeplot(dataf):
dataf = dataf.copy()
plt_data = dataf["ml"]
plt_data = prep_sizebplot(plt_data)
plt = sns.barplot(x='year', y='size', data=plt_data, palette="Blues_d")
plt.set_title('Number of households in simulation')
fig = plt.get_figure()
return fig
def countplot(dataf):
dataf = dataf.copy()
orig_df = dataf["standard"]
orig_data = orig_df[orig_df['predicted'] == 0]
plt_data = prep_countbplot(orig_data)
plt = sns.barplot(x='year', y='relative', data=plt_data, palette="Blues_d")
fig = plt.get_figure()
return fig
def plot_freqs(files, new_route="./data",
old_route="./not_treated_data", freqs=[], times=[], errors=[]):
for filename in files:
sheets = pd.read_excel(old_route + "/" + filename,
sheet_name=None)
sheet_name, df = next(iter(sheets.items()))
df = df.drop(df.columns[0], axis=1)
print(df)
df = df *30
print(df)
freq = df.to_numpy().sum() / len(df.columns)
error = df.to_numpy().std()
errors.append(error)
freqs.append(freq)
times.append(filename.split("_")[-1].split(".")[0])
if sorted(times) == ["April", "August", "December", "February",
"January", "July", "June", "March", "May", "November",
"October", "September"]:
months = ["January", "February", "March", "April", "May",
"June", "July", "August", "September", "October",
"November", "December"]
for i in range(len(times)):
for j in range(len(months)):
if times[i] == months[j]:
times[i] = j+1
freqs = [x for y, x in sorted(zip(times,freqs))]
errors = [x for y, x in sorted(zip(times,errors))]
times = months
else: ## if we do not have to order by months, we order by names
freqs = [x for y, x in sorted(zip(times,freqs))]
errors = [x for y, x in sorted(zip(times,errors))]
times = sorted(times)
df = pd.DataFrame([times, freqs, errors],
index=["Times", "Frequencies", "Deviations"],
columns=range(len(times)))
new_df = df.T
maxy = roundup2(max(freqs) + max(errors),
pos=len(str(ceil(max(freqs))))-1)
miny = 0
ticks = [miny, (miny + maxy)/3, 2*(miny + maxy)/3, maxy]
plt = new_df.plot(kind="bar", x="Times", y="Frequencies",
yerr="Deviations", yticks=ticks, legend=False, figsize=(10,10),
capsize=5, fontsize=10)
plt.tick_params(axis="x", labelsize=10, rotation=30)
plt.set_xlabel("Months")
plt.set_ylabel("Absolute Frequency")
plt.set_title("Histogram of monthly absolute frequencies")
scientific_formatter = FuncFormatter(scientific)
plt.yaxis.set_major_formatter(scientific_formatter)
fig = plt.get_figure()
fig.savefig(new_route + "/Freqs_vs_Time.png")
def getPlot(self, params):
ticke= str(params['ticke'])
ticker3 = str(params['ticker3'])
year = float(params['year'])
df = pd.read_csv('provinces/vhi_id_%s.csv'%ticker3,index_col=False, header=1)
if year!=0:
w=df[df['year']==year]
else:
w=df
x=w[[ticke]]
y =w['week']
plt = x.plot(y)
fig = plt.get_figure()
return fig
def histogram():
ten = pd.to_numeric(pivot_cost['2010']).dropna()
eleven = pd.to_numeric(pivot_cost['2011']).dropna()
twelve = pd.to_numeric(pivot_cost['2012']).dropna()
thirteen = pd.to_numeric(pivot_cost['2013']).dropna()
fourteen = pd.to_numeric(pivot_cost['2014']).dropna()
plt = sns.distplot(fourteen)
plt1 = sns.distplot(thirteen)
plt2 = sns.distplot(twelve)
plt3 = sns.distplot(eleven)
plt4 = sns.distplot(ten)
fig = plt.get_figure()
fig.savefig("overlay.png")
colors = [
plt.cm.Spectral(each) for each in np.linspace(0, 1, len(unique_labels))
]
for k, col in zip(unique_labels, colors):
if k == -1:
# Black used for noise.
col = [0, 0, 0, 1]
class_member_mask = (labels == k)
xy = X[class_member_mask & core_samples_mask]
plt.plot(xy[:, 0],
xy[:, 1],
'o',
markerfacecolor=tuple(col),
markeredgecolor='k',
markersize=14)
xy = X[class_member_mask & ~core_samples_mask]
plt.plot(xy[:, 0],
xy[:, 1],
'o',
markerfacecolor=tuple(col),
markeredgecolor='k',
markersize=6)
plt.title('Estimated number of clusters: %d' % n_clusters_)
plt.show()
plt.get_figure().savefig('dbscan.png', bbox_inches='tight')
epochs=10,
verbose=1,
validation_data=(x_test, y_test))
model.summary()
#Create bar chart with train and val accuracies
train_accuracy = fit_info.history['accuracy']
test_accuracy = fit_info.history['val_accuracy']
df = pd.DataFrame(columns = ['train_accuracy'], index = range (1,epochs+1), data=train_accuracy)
df['test_accuracy'] = test_accuracy
plt = df.plot.bar(rot = 1, figsize = (12,10))
plt.grid(axis = 'y')
plt.set_ylim(0.8, 1)
fig = plt.get_figure()
fig.savefig("accuracy_plot.png")
print(df['test_accuracy'].max())
## Looping through different regularization factors and creating replicates to find the optimal parameter value (based on accuracy)
## New amount of units in layers: 300 & 500
from tensorflow.keras import regularizers
model = Sequential()
epochs = 40
regularization_factor = [0.000001,0.00025,0.00050,0.00075,0.0010]
max_accuracies = []
factor_count = 0
def draw_plot(filename):
ds = pd.read_csv(filename, index_col=['Date'], parse_dates=['Date'])
ds.as_matrix()
plt = ds.plot(figsize=(12, 6))
plt.get_figure().savefig("files/" + str(os.path.splitext(os.path.basename(filename))[0]) + "_origin.png")
def show_mappings(Z, path):
plt = sns.scatterplot(Z[:, 0], Z[:, 1], hue=Z[:, 2])
fig = plt.get_figure()
fig.savefig(path)
