import altair as alt
import altair_viewer
driver = GraphDatabase.driver("bolt://localhost", auth=("neo4j", "neo"))
with driver.session(database="foo") as session:
result = session.run("""
MATCH (p:Place)
RETURN p.name AS place, p.node2vecMoreTraining AS embedding, p.community AS community, p.country AS country
""")
X = {row["place"]: {"embedding": row["embedding"], "community": row["community"], "country": row["country"]} for row in result}
X_embedded = TSNE(n_components=2, random_state=6).fit_transform([X[key]["embedding"] for key in X.keys()])
places = list(X.keys())
df = pd.DataFrame(data = {
"place": places,
"community": [f"Community-{X[place]['community']}" for place in places],
"country": [X[place]['country'] for place in places],
"x": [value[0] for value in list(X_embedded)],
"y": [value[1] for value in list(X_embedded)]
})
chart = alt.Chart(df).mark_circle(size=60).encode(
x='x',
y='y',
color='community',
tooltip=['place', 'community', 'country']
).properties(width=800, height=500)
altair_viewer.display(chart)
def plot_reported_cases(county='Alle_Fylker',
periodicity='day',
start=None,
end=None,
plot_lib='Altair',
view_plot=False,
plot_title=False,
width=800,
height=500):
"""
Function to generate a bar-plot for Covid19's total daily cases from
FHI/Norway.
Parameters:
~~~~~~~~~~~~~~~~~~~~
:param county: str: Name od the county to be analyzed, which could
be 'Agder', 'Innlandet', 'More_og_Romsdal',
'Nordland', 'Oslo', 'Rogaland',
'Troms_og_Finnmark', Trondelag',
'Vestfold_og_Telemark', 'Vestland',
'Viken'or 'Alle_Fylker' for all counties.
:param periodicity: str: Periodicity to display the plot, which could
be 'day' or 'week'. Default 'day'.
:param start: str or date object: The starting date for the analysis.
:param end: str or date object: The ending date for the analysis.
:param plot_lib: str: The name of the plotting library to be used.
Default is 'Altair' which uses Altair. A option
is 'plt' which uses matplotlib.
:param view_plot: bool: The default is False to not show the plot.
:param plot_title: bool: The default is False to not display the
plot's title.
:param width: int: Only used for 'Altair' plot. Define the width of
the plotting. Default is 800.
:param height: int: Only used for 'Altair' plot. Define the height of
the plotting. Default is 500.
Notes:
~~~~~~~~~~~~~~~~~~~~
- If start and end are None, them it will plot the entire data-base.
"""
# setting fixed parameters
col = 'Nye tilfeller'
title = 'Number of reported cases'
# setting prefix and suffix for the data-files
path = 'data/'
if periodicity == 'day':
prefix = path + 'antall-meldte-covid-19-t_day_'
elif periodicity == 'week':
prefix = path + 'antall-meldte-covid-19-t_week_'
else:
raise ValueError("Error: The given periodicity is not valid.")
suffix = '.csv'
# reading data as DataFrame
df = pd.read_csv(filepath_or_buffer=prefix + county + suffix,
sep=',',
index_col='Dato')
# setting data range
df = df[start:end]
# creating Altair's plotting
if plot_lib == 'Altair':
# fixing datetime for Altair
df.reset_index(inplace=True)
if periodicity == 'day':
df['Dato'] = pd.to_datetime(df["Dato"], dayfirst=True)
# plotting bar and lines in dual-axis
chart = alt.Chart(df).mark_bar().encode(alt.X('Dato:T'),
alt.Y(col + ':Q',
title=title),
tooltip=['Dato', col])
elif periodicity == 'week':
# plotting bar and lines in dual-axis
chart = alt.Chart(df).mark_bar().encode(alt.X(
'Dato', scale=alt.Scale(zero=False)),
alt.Y(col + ':Q',
title=title),
tooltip=['Dato', col])
# plotting
chart = chart.properties(
width=width,
height=height,
title=f'Data Vs ({title}) for {county}').interactive()
# viewing the plot
if view_plot is True:
# altair_viewer.show(chart)
altair_viewer.display(chart)
return chart
# creating Matplotlib plot
elif plot_lib == 'plt':
# plotting using matplotlib
plt.bar(x=df.index, height=df[col])
# plotting set-up
if plot_title is True:
plt.title(f'Data Vs {title} for {county}')
plt.xlabel('Date')
plt.ylabel(f'{title}')
plt.xticks(rotation='vertical')
# viewing plot
if view_plot is True:
plt.show()
else:
raise ValueError(f"Error: Plot type {plot_lib} not implemented.")
data = data_new.copy()
mode = 'rf'
# assign to variables
X = data.drop(columns=['class'])
y = data['class']
# encoding: Label encoding for binary class, one-hot encoding for the nominal variables
y = sklearn.preprocessing.LabelEncoder().fit_transform(y)
X = pd.get_dummies(X)
log_reg = sklearn.linear_model.LogisticRegression(max_iter=10000)
lda = sklearn.discriminant_analysis.LinearDiscriminantAnalysis()
gnb = sklearn.naive_bayes.GaussianNB()
models = [log_reg, lda, gnb]
X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(
X, y, test_size=0.2, random_state=1)
# threshold charts
threshold_chart = get_score_threshold_chart(X_train, X_test, y_train,
y_test, 'nb', 'F2')
# roc charts
# chart_secondary = get_roc_chart(X_train, X_test, y_train, y_test, mode)
# chart_original = get_roc_chart(X_train, X_test, y_train, y_test, mode)
""" comment/uncomment depending on the chart that shall be displayed"""
chart = get_roc_chart(X_train, X_test, y_train, y_test, mode)
chart.save('roc.html')
view.display(chart)
#view.display(get_balance_chart(data_new, title='Balance plot for secondary data', reindex=['p', 'e']))
#view.display(get_balance_chart(data_original, title='Balance plot for 1987 data'))#, count=False, reindex=['p', 'e']))
#view.display(get_correlation_chart(data_new, show_progress=True))
#view.display(get_correlation_chart(data_original, show_progress=True))
#view.display(threshold_chart)
def candlestick(source, width=800, height=500, view=True, lines=None):
"""
Function to generate a interactive candlestick chart for visualization of
the time series of a financial source.
Parameters:
===================
:param source: pd.DataFrame: Time series DataFrame containing
OLHCV values.
:param width: int: The width of the chart.
:param height: int: The height of the chart.
:param view: bool: If True, it will return a URL to visualize the chart.
:param lines: dict: Containing as keys the name of the columns and as
values the colors of the lines.
Return:
===================
The function returns a URL where the interactive chart will be displayed.
"""
source.reset_index(inplace=True)
# defining colors for the candlesticks
open_close_color = alt.condition("datum.open <= datum.close",
alt.value("#06982d"),
alt.value("#ae1325"))
# creating the base for the candlestick's chart
base = alt.Chart(source).encode(
alt.X('index:T',
axis=alt.Axis(format='%Y/%m/%d', labelAngle=-90, title='Dates')),
color=open_close_color,
)
# creating a line for highest and lowest
rule = base.mark_rule().encode(
alt.Y(
'low:Q',
scale=alt.Scale(zero=False),
axis=alt.Axis(title='Prices', orient='right'),
), alt.Y2('high:Q'))
# creating the candlestick's bars
bar = base.mark_bar().encode(alt.Y('open:Q'), alt.Y2('close:Q'))
# joining OLHC together
chart = rule + bar
# drawing line
# !!!!!!!!!! need to fix the problem with the colors
if lines is not None:
for k, v in lines.items():
chart += base.mark_line(color=v, opacity=0.3).encode(y=alt.Y(k))
# adding tooltips, properties and interaction
chart = chart.encode(tooltip=[
alt.Tooltip('index:T', title='Date'),
alt.Tooltip('open', title='Open'),
alt.Tooltip('low', title='Low'),
alt.Tooltip('high', title='High'),
alt.Tooltip('close', title='Close'),
alt.Tooltip('volume', title='Volume')
]).properties(width=width,
height=height,
title=f'Candlestick visualization').interactive()
# creating x-axis selections
# !!!!!!!!!! it is jumping bar - fix later
nearest = alt.selection(type='single',
nearest=True,
on='mouseover',
fields=['index'],
empty='none')
# drawing a vertical rule at the location of the selection
v_rule = alt.Chart(source).mark_rule(color='gray').encode(
x='index:T', ).transform_filter(nearest)
# adding nearest selection on candlestick's chart
chart = chart.add_selection(nearest)
# ##########
if view is True:
# altair_viewer.show(chart)
altair_viewer.display(chart + v_rule)