def pl(df, r, var):
tmp = df[df.randomSeed.isin(r)]
plot(px.line(tmp, height=300 * len(r), x="tick",
y = var,
color="FirmNumID",
line_dash="scenario",
facet_row="randomSeed"
))
def plMelt(df, r, vars, id_vars=["randomSeed","scenario","tick","FirmNumID"]):
tmp = df[df.randomSeed.isin(r)]
tmp = tmp.melt(id_vars=id_vars, value_vars=vars)
plot(px.line(tmp, height=300 * len(r), x="tick",
y= "value",
color="FirmNumID",
line_dash="scenario",
facet_col="variable",
facet_row="randomSeed"
))
def plot_subr_line_cat(filtered_json_week, nperc):
filtered_df_week = pd.read_json(filtered_json_week, orient='split')
filtered_df_week_cat = get_flag_perc(filtered_df_week,
['subreddit_cat', 'week'])
subr_line_cat = px.line(filtered_df_week_cat,
x="week",
y="ISFLAG" if nperc == "n" else "perc_FLAG",
color="subreddit_cat")
for l in subr_line_cat.data:
l.update(mode='markers+lines')
return subr_line_cat
def update_figure(selected_year):
filtered_lc = lc[lc['issue_year'] <= selected_year]
return px.line(filtered_lc.groupby(['issue_d', 'term'
]).mean()[['int_rate']].reset_index(),
x='issue_d',
y='int_rate',
color='term',
labels={
'issue_d': 'Issue Date',
'int_rate': 'Interest Rate'
})
def update_figure(selected_year):
filtered_lc = lc[lc['issue_year'] <= selected_year]
return px.line(filtered_lc.groupby(['issue_d', 'term'
]).sum()[['loan_amnt']].reset_index(),
x='issue_d',
y='loan_amnt',
labels={
'issue_d': 'Issue Date',
'loan_amnt': 'Origination Principal'
},
color='term')
def plot_subr_line_20(filtered_json_week, nperc):
filtered_df_week = pd.read_json(filtered_json_week, orient='split')
filtered_df_week_20 = get_flag_perc(
filtered_df_week[filtered_df_week.subreddit.isin(top20subreddits)],
groups=['subreddit', 'week'])
subr_line_20 = px.line(filtered_df_week_20,
x="week",
y="ISFLAG" if nperc == "n" else "perc_FLAG",
color="subreddit")
for l in subr_line_20.data:
l.update(mode='markers+lines')
return subr_line_20
def update_hourly_graph(query, selected_label_id, min_prob, max_prob,
selected_iteration_id, clickData):
try:
query_dict = parse_qs(query[1:])
project_id = query_dict['project'][0]
selected_date = clickData['points'][0]['x']
with dashapp.server.app_context():
per_file = db.session.query(
ModelOutput.file_name,
func.count(ModelOutput.id).label('count')).filter(
ModelOutput.iteration_id == selected_iteration_id,
ModelOutput.label_id == selected_label_id,
ModelOutput.probability >= float(min_prob),
ModelOutput.probability <= float(max_prob)).group_by(
ModelOutput.file_name).subquery()
per_hour = db.session.query(
MonitoringStation.name.label('station'),
func.sum(per_file.columns.count).label('count'),
func.date(AudioFile.timestamp).label('date'),
func.extract('hour', AudioFile.timestamp).label('hour')
).join(
AudioFile,
per_file.columns.file_name == AudioFile.name).join(
Equipment,
AudioFile.sn == Equipment.serial_number).join(
MonitoringStation).group_by('station').group_by(
'date').group_by('hour').filter(
MonitoringStation.project_id == project_id,
func.date(AudioFile.timestamp) ==
selected_date).order_by('hour')
df = pd.read_sql(per_hour.statement, db.session.bind)
fig = px.line(df,
x='hour',
y='count',
color='station',
title='Hourly count for ' + selected_date)
fig.update_traces(mode='markers')
except:
fig = px.line()
return fig
def update_hourly_graph(query, selected_label_id, min_prob, max_prob,
selected_iteration_id, clickData):
try:
project_id = get_project_id(query)
selected_date = clickData['points'][0]['x']
with dashapp.server.app_context():
per_file = db.session.query(
ModelOutput.file_name,
func.count(ModelOutput.id).label('count')).filter(
ModelOutput.iteration_id == selected_iteration_id,
ModelOutput.label_id == selected_label_id,
ModelOutput.probability >= float(min_prob),
ModelOutput.probability <= float(max_prob)).group_by(
ModelOutput.file_name).subquery()
per_hour = db.session.query(
func.min(MonitoringStation.id).label('station_id'),
MonitoringStation.name.label('station'),
func.sum(per_file.columns.count).label('count'),
func.date(AudioFile.timestamp).label('date'),
func.extract(
'hour', AudioFile.timestamp).label('hour')).join(
AudioFile,
per_file.columns.file_name == AudioFile.name).join(
AudioFile.monitoring_station
).group_by('station').group_by('date').group_by(
'hour').filter(
MonitoringStation.project_id == project_id,
func.date(AudioFile.timestamp) ==
selected_date).order_by('hour')
df = pd.read_sql(per_hour.statement, db.session.bind)
station_colors = get_station_colors(project_id)
label = Label.query.get(selected_label_id)
fig = px.line(df,
x='hour',
y='count',
custom_data=['station_id', 'station'],
color='station',
color_discrete_map=station_colors,
title='Hourly count for {label} on {date}'.format(
label=label, date=selected_date))
fig.update_traces(mode='markers')
except:
fig = go.Figure()
fig.update_layout(title='Hourly count',
annotations=[{
'text': 'Select a day',
'showarrow': False,
'font': {
'size': 28
}
}])
return fig
def global_plot_create(data, x, y, title, xaxis, yaxis):
fig = px.line(data, x=x, y=y, color='Country/Region', width=1000, height=800) # we assign the size of the plot, assigning the width and the height.
fig.update_layout(title=title, # in this part we assign the layout to include the xaxis and yaxis
xaxis_title= xaxis,
yaxis_title = yaxis,
legend_title_text='Countries',
yaxis_type="log",
yaxis_tickformat = 'f',
xaxis_gridcolor = 'LightGreen', # the grid will make the line that appears on graph to be more flexible to be studied
yaxis_gridcolor = 'LightGreen',
paper_bgcolor='rgba(0,0,0,0)',
plot_bgcolor='rgba(0,0,0,0)')
return fig
def us_plot_create(data, x, y, title, xaxis, yaxis):
fig = px.line(data, x=x, y=y, color='Province_State', width=1000, height=800)
fig.update_layout(title=title,
xaxis_title= xaxis,
yaxis_title = yaxis,
legend_title_text='United Kingdom',
yaxis_type="log",
yaxis_tickformat = 'f',
xaxis_gridcolor = 'LightGreen',
yaxis_gridcolor = 'LightGreen',
paper_bgcolor='rgba(0,0,0,0)',
plot_bgcolor='rgba(0,0,0,0)')
return fig
def update_graph(query, selected_label_id, min_prob, max_prob, start_hour,
end_hour, selected_iteration_id):
try:
if selected_label_id == None:
raise
project_id = get_project_id(query)
with dashapp.server.app_context():
per_file = db.session.query(
ModelOutput.file_name,
func.count(ModelOutput.id).label('count')).filter(
ModelOutput.iteration_id == selected_iteration_id,
ModelOutput.label_id == selected_label_id,
ModelOutput.probability >= float(min_prob),
ModelOutput.probability <= float(max_prob)).group_by(
ModelOutput.file_name).subquery()
per_day = db.session.query(
func.min(MonitoringStation.id).label('station_id'),
MonitoringStation.name.label('station'),
func.sum(per_file.columns.count).label('count'),
func.date(AudioFile.timestamp).label('date')).join(
AudioFile, per_file.columns.file_name == AudioFile.name
).join(AudioFile.monitoring_station).filter(
MonitoringStation.project_id == project_id,
func.extract('hour', AudioFile.timestamp) >=
start_hour,
func.extract('hour', AudioFile.timestamp) <= end_hour,
).group_by('station').group_by('date').order_by('date')
df = pd.read_sql(per_day.statement, db.session.bind)
station_colors = get_station_colors(project_id)
label = Label.query.get(selected_label_id)
fig = px.line(
df,
x='date',
y='count',
custom_data=['station_id', 'station'],
color='station',
color_discrete_map=station_colors,
title='Daily count for {label} {start}:00-{end}:59'.format(
label=label, start=start_hour, end=end_hour))
fig.update_traces(mode='markers')
except:
fig = go.Figure()
fig.update_layout(title='Daily count',
annotations=[{
'text': 'Please select a label',
'showarrow': False,
'font': {
'size': 28
}
}])
return fig
def linechar(chosed_asin, table_chosed_asin, asin_month_rank):
fig = px.line(table_chosed_asin,
x=table_chosed_asin['date'],
y=table_chosed_asin[f'{chosed_asin}'],
title=f'{chosed_asin}的当月排名情况(红线为当月排名)')
fig.add_shape(type='line',
x0=table_chosed_asin['date'].iloc[0],
y0=asin_month_rank,
x1=table_chosed_asin['date'].iloc[-1],
y1=asin_month_rank,
line=dict(color='Red'),
xref='x',
yref='y')
return fig
def showPlots(chart_select):
if chart_select == 'Scatterplots':
st.sidebar.subheader("Scatterplot Settings")
try:
x_values = st.sidebar.selectbox('X axis', options=numeric_columns)
y_values = st.sidebar.selectbox('Y axis', options=numeric_columns)
color_value = st.sidebar.selectbox("Color", options=non_numeric_columns)
plot = px.scatter(data_frame=df, x=x_values, y=y_values, color=color_value)
# display the chart
st.plotly_chart(plot)
except Exception as e:
print(e)
if chart_select == 'Lineplots':
st.sidebar.subheader("Line Plot Settings")
try:
x_values = st.sidebar.selectbox('X axis', options=numeric_columns)
y_values = st.sidebar.selectbox('Y axis', options=numeric_columns)
color_value = st.sidebar.selectbox("Color", options=non_numeric_columns)
plot = px.line(data_frame=df, x=x_values, y=y_values, color=color_value)
st.plotly_chart(plot)
except Exception as e:
print(e)
if chart_select == 'Histogram':
st.sidebar.subheader("Histogram Settings")
try:
x = st.sidebar.selectbox('Feature', options=numeric_columns)
bin_size = st.sidebar.slider("Number of Bins", min_value=10,
max_value=100, value=40)
color_value = st.sidebar.selectbox("Color", options=non_numeric_columns)
plot = px.histogram(x=x, data_frame=df, color=color_value)
st.plotly_chart(plot)
except Exception as e:
print(e)
if chart_select == 'Boxplot':
st.sidebar.subheader("Boxplot Settings")
try:
y = st.sidebar.selectbox("Y axis", options=numeric_columns)
x = st.sidebar.selectbox("X axis", options=non_numeric_columns)
color_value = st.sidebar.selectbox("Color", options=non_numeric_columns)
plot = px.box(data_frame=df, y=y, x=x, color=color_value)
st.plotly_chart(plot)
except Exception as e:
print(e)
def update_graph(user_id):
# select from condition table
# progression data
query = 'select * from user_ALSFRS_score where user_id = ' + str(user_id)
user_progression = pd.read_sql(query, con=engine)
user_progression['report_date'] = pd.to_datetime(
user_progression['report_date'], format='%Y-%m-%d')
user_progression['report_date'] = user_progression['report_date'].map(
dt.datetime.toordinal)
user_progression = user_progression.sort_values(by='report_date')
time = user_progression['report_date'].to_numpy()
score = user_progression['score'].to_numpy()
# condition data
query = 'select * from user_condition where user_id = ' + str(user_id)
user_conditions = pd.read_sql(query, con=engine)
# generate user condition code to predict progression speed
# and get true user progression speed
user_conditions_code = np.asarray(
fun.getConditions(CONDITION_MAP, user_conditions))
# pedict progression speed
progression_speed_predicted = reg_progression.predict(
user_conditions_code.reshape(1, -1))
predicted_score = progression_speed_predicted[0][
0] * time + progression_speed_predicted[0][1]
# shift up or down based on subject initial score
predicted_score = predicted_score - (predicted_score[0] - score[0])
df_predicted = pd.DataFrame({
'report_date': time,
'score': predicted_score
})
# begin plot
#true_m, true_b = np.polyfit(time, score, 1)
user_progression['report_date'] = user_progression['report_date'].map(
dt.datetime.fromordinal)
df_predicted['report_date'] = df_predicted['report_date'].map(
dt.datetime.fromordinal)
figure = px.scatter(user_progression,
x="report_date",
y="score",
trendline="ols")
figure.update_traces(name='Actural user data', showlegend=True)
figure2 = px.line(df_predicted, x='report_date', y='score')
figure2.update_traces(name='Predicted progression', showlegend=True)
figure2.update_traces(line_color='#147852')
figure.add_trace(figure2.data[0])
return figure
def summary_plot2(n_clicks, bank2, y2, rprt_date2):
plot_data = data_source[(data_source['RSSD'].isin(bank2))
& (data_source['date_index'] >= rprt_date2[0]) &
(data_source['date_index'] <= rprt_date2[1])]
sum_plot = px.line(
plot_data,
x="Report_Date",
y=y2,
color="bank",
hover_name="bank",
line_shape="spline",
labels={
'avg_os': 'Average Loan Balance (in ths)',
'alll': 'Loan Loss Reserve (in ths)',
'alll_rate': 'Loan Loss Reserve (in %)',
'ncl': 'Net Credit Loss (in ths)',
'ncl_rate': 'Net Credit Loss (in %)',
'alll_to_ncl': 'ALLL / NCL Ratio',
'alll_to_nonaccrual': 'ALLL / Non-accrual Ratio'
},
template='plotly_dark',
)
sum_plot['layout'].update(
{
'plot_bgcolor': 'rgba(0,0,0,0)',
'paper_bgcolor': 'rgba(0,0,0,0)',
},
xaxis=dict(showgrid=True,
gridcolor='#747480',
gridwidth=0.5,
zerolinecolor='#747480',
zerolinewidth=1,
linecolor='#FFE600',
linewidth=2),
yaxis=dict(showgrid=True,
gridcolor='#747480',
gridwidth=0.5,
zerolinecolor='#747480',
zerolinewidth=1,
linecolor='#FFE600',
linewidth=2))
return sum_plot
def plot_subr_line_cat(filtered_json_week, nperc):
filtered_df_week = pd.read_json(filtered_json_week, orient='split')
filtered_df_week_cat = get_flag_perc(filtered_df_week,
['subreddit_cat', 'week'])
subr_line_cat = px.line(filtered_df_week_cat,
x="week",
y="ISFLAG" if nperc == "n" else "perc_FLAG",
color="subreddit_cat")
for l in subr_line_cat.data:
l.update(mode='markers+lines')
tickvals0 = get_xlabs((min([min(a.x) for a in subr_line_cat.data]),
max([max(a.x) for a in subr_line_cat.data])))
subr_line_cat.update_layout(
xaxis=dict(tickmode='array',
tickvals=tickvals0,
ticktext=week_day_df['label'][tickvals0]))
return subr_line_cat
def update_detailed_file_graph(query, file_click_data, daily_click_data):
try:
project_id = get_project_id(query)
if callback_context.triggered[0][
'prop_id'] == 'file-graph.clickData':
selected_date = file_click_data['points'][0]['x']
if callback_context.triggered[0][
'prop_id'] == 'daily-graph.clickData':
selected_date = daily_click_data['points'][0]['x']
with dashapp.server.app_context():
project_files = AudioFile.query.join(
AudioFile.monitoring_station).filter(
MonitoringStation.project_id == project_id,
func.date(AudioFile.timestamp) == selected_date
).with_entities(
AudioFile.size, AudioFile.timestamp,
MonitoringStation.name.label('station')).subquery()
bytes_per_hour = db.session.query(
func.sum(project_files.c.size).label('bytes'),
func.extract('hour',
project_files.c.timestamp).label('hour'),
project_files.c.station).group_by('station').group_by(
'hour').order_by('hour')
df = pd.read_sql(bytes_per_hour.statement, db.session.bind)
df['MB'] = df['bytes'] / 1024 / 1024
station_colors = get_station_colors(project_id)
fig = px.line(df,
x='hour',
y='MB',
color='station',
color_discrete_map=station_colors,
title='Data recorded per hour on ' + selected_date)
fig.update_traces(mode='markers')
fig.update_layout(hovermode='x')
except:
fig = go.Figure()
fig.update_layout(title='Data recorded per hour',
annotations=[{
'text': 'Select a day',
'showarrow': False,
'font': {
'size': 28
}
}])
return fig
def plot_tweets_time(df):
'''
Plots tweets by week for each unique handle
'''
df_weekly_count = df.groupby(['date_week', 'handle'],
as_index=False).count().iloc[:, 0:3]
df_weekly_count.columns = ['week', 'handle', 'number of tweets']
# create plot
fig = px.line(df_weekly_count,
x='week',
y='number of tweets',
color="handle",
color_discrete_map=colour_dict,
title="Number of Tweets by Week",
height=400)
fig.update_layout({"showlegend": False})
fig.update_layout(margin=dict(l=0, r=0, t=30, b=30))
return fig
def global_plot_create(data, x, y, title, xaxis, yaxis):
fig = px.line(data,
x=x,
y=y,
color='Country/Region',
width=800,
height=600)
fig.update_layout(
title=title,
xaxis_title=xaxis,
yaxis_title=yaxis,
legend_title_text='Countries',
#yaxis_type="log",
yaxis_tickformat='f',
xaxis_gridcolor='LightBlue',
yaxis_gridcolor='LightBlue',
paper_bgcolor='rgba(0,0,0,0)',
plot_bgcolor='rgba(0,0,0,0)')
return fig
def comisiones_por_mes(df):
df_productos = df.loc[:, [
"pk_cid",
"pk_partition",
] + PRODUCTS]
df_ventas_mes = df_productos[df_productos[PRODUCTS].sum(
axis=1) >= 1].pivot_table(values=PRODUCTS,
index=['pk_cid', 'pk_partition'])
df_ventas_mes.reset_index(inplace=True)
df_ventas_mes.set_index('pk_partition', inplace=True)
df_ventas_mes.drop_duplicates(keep='first', inplace=True)
df_ventas_producto_mes = df_ventas_mes.groupby(
by=df_ventas_mes.index).sum()
del (df_ventas_producto_mes['pk_cid'])
df_ventas_producto_mes = df_ventas_producto_mes[3:]
fig = px.line(df_ventas_producto_mes,
x=df_ventas_producto_mes.index,
y=PRODUCTS,
title='Evolución de ventas por producto')
return fig, df_ventas_producto_mes
def activos_web(df_clientes):
df_numero_clientes_activos = df_clientes[
df_clientes["active_customer"] == 1].groupby(by=["pk_partition"])[[
"pk_cid"
]].count().sort_values(by="pk_partition", ascending=False)
df_numero_clientes = df_clientes.groupby(by=["pk_partition"])[[
"pk_cid"
]].count().sort_values(by="pk_partition", ascending=False)
df_evolucion_clientes_activos_web = pd.concat(
[df_numero_clientes, df_numero_clientes_activos], axis=1)
df_evolucion_clientes_activos_web.columns = ["total", "activos_web"]
df_evolucion_clientes_activos_web["porcentaje"] = round(
df_evolucion_clientes_activos_web["activos_web"] /
df_evolucion_clientes_activos_web["total"] * 100, 2)
df_evolucion_clientes_activos_web.reset_index(inplace=True)
fig = px.line(df_evolucion_clientes_activos_web,
x="pk_partition",
y="porcentaje",
range_y=[0, 50],
title='Evolución tasa de navegación en la web')
return fig, df_evolucion_clientes_activos_web
def tsa_figrue(x,y,color,facet_col,facet_row,summarizer='mean'):
agg_cols = []
groups = [x,color,facet_col,facet_row]
for var in groups:
if((var is not None) and (var in tsa_df.columns)):
agg_cols.append(var)
graph_df = tsa_df.copy()
df = graph_df.groupby(agg_cols).agg({y:summarizer}).reset_index()
fig = px.line(
df,
x,
y,
color=color,
facet_col=facet_col,
facet_row=facet_row,
facet_col_wrap=4
)
if(not(facet_col is None) or not(facet_row is None)):
fig.for_each_annotation(lambda a: a.update(text=a.text.split("=")[1]))
return fig
def line_chart(df, option):
g = px.line(
df,
x='Date',
y='Price',
color='Winner',
# animation_frame='Date',
color_discrete_sequence=[
'#FF1493', '#120052', '#652EC7', '#00C2BA', '#82E0BF', '#55E0FF'
],
title='Betting Odds Over Time')
g.update_traces(mode='lines', opacity=.75, line=dict(width=4))
g.update_yaxes(title='Implied Probability',
showgrid=True,
gridwidth=1,
gridcolor='#D4D4D4')
g.update_layout(plot_bgcolor='white')
g.update_xaxes(title='Date',
showgrid=False,
gridwidth=1,
gridcolor='#D4D4D4')
st.plotly_chart(g, use_container_width=True)
def update_trace(value_cam_ID, value_time):
if value_time == 'Hour':
query = 'SELECT time, cam_name, cam_ID, AVG(vehicles) as average_vehicles FROM traffic_cams WHERE time >= NOW() - INTERVAL 1 HOUR and mod(minute(time),5) = 0 GROUP BY CONCAT(cam_ID, time)'
if value_time == 'Day':
query = 'SELECT time, cam_name, cam_ID, AVG(vehicles) as average_vehicles FROM traffic_cams WHERE time >= NOW() - INTERVAL 24 HOUR and mod(minute(time),120) = 0 GROUP BY CONCAT(cam_ID, time)'
if value_time == 'Week':
query = 'SELECT time, cam_name, cam_ID, AVG(vehicles) as average_vehicles FROM traffic_cams WHERE time >= NOW() - INTERVAL 168 HOUR and mod(minute(time),840) = 0 GROUP BY CONCAT(cam_ID, time)'
df = pd.read_sql(query, con=engine)
df['time_by_m'] = df['time'].dt.floor('1Min')
df = df.drop_duplicates(['time_by_m', 'cam_ID'])
df['time_mark'] = df['time_by_m'].astype(str)
df.sort_values(by='time')
df2 = df.loc[df['cam_ID'].isin(value_cam_ID)]
figure = px.line(df2,
x='time_mark',
y="average_vehicles",
color="cam_name",
line_shape='spline')
return figure
def update_graph_live(n):
global df_global_saison
global df_global_cc
if (n % 150 == 0):
ryzom = weather(True)
else:
ryzom = weather(False)
df_global_saison = ryzom.getSeason()
df = pd.DataFrame(ryzom.getData())
df_dates = df.loc[df['dates']==datetime.now().strftime("%H:%M"),:]
df_global_cc = get_CC_libelle(df_dates.iloc[0,2])
fig = px.line (df,x = "dates", y = "cc" ,color = "continents", title= "Saison : " + df_global_saison + " conditions climatiques : " + df_global_cc)
fig.update_yaxes(tickmode = "array",tickvals = [0,1670,3340,5000,6666,8340,10000],ticktext = ["Best", "Good 16.7%","Good 33.4%","Bad 50%","Bad 66.6%","Worst 83.4%","Worst"],showgrid = True)
fig.update_xaxes(tickmode = "linear",showline = True,showgrid = True, dtick = 6)
fig.update_layout(legend_uirevision='true')
fig.add_vline(x=datetime.now().strftime("%H:%M"), line_width=1, line_dash="dash", line_color="red")
fig.update_xaxes(showline=True, linewidth=2, linecolor='black', gridcolor='rgba(211, 211, 211, 0.75)')
#fig.update_layout({'plot_bgcolor': 'rgba(0, 0, 0, 0)','paper_bgcolor': 'rgba(0, 0, 0, 0)'})
#fig.update_yaxes(showline=True, linewidth=2, linecolor='black', gridcolor='grey')
return fig
def calcula_ingresos_por_mes(df):
df_ventas_producto_mes = comisiones_por_mes(df)[1]
df_ingresos = pd.DataFrame()
l_ahorro = ['long_term_deposit', 'pension_plan', 'short_term_deposit']
l_inversion = ['funds', 'securities']
l_financiacion = ['credit_card', 'loans', 'mortgage']
l_cuenta = [
'debit_card', 'em_account_p', 'em_account_pp', 'em_acount',
'emc_account', 'payroll', 'payroll_account'
]
for c in df_ventas_producto_mes.columns:
if c in l_ahorro or c in l_inversion:
df_ingresos[c] = df_ventas_producto_mes[c] * 40
elif c in l_financiacion:
df_ingresos[c] = df_ventas_producto_mes[c] * 60
else:
df_ingresos[c] = df_ventas_producto_mes[c] * 10
fig = px.line(df_ingresos,
x=df_ingresos.index,
y=PRODUCTS,
title='Evolución de ingresos por producto')
return fig, df_ingresos
def evolucion_clientes(df_clientes):
_df = df_clientes[df_clientes[PRODUCTS].sum(axis=1) >= 1].groupby(
by=["pk_partition"])[["pk_cid"]].count().sort_values(by="pk_partition",
ascending=False)
_df["ant"] = _df["pk_cid"].shift(-1)
_df["dif"] = _df["pk_cid"] - _df["ant"]
df_n_cli = df_clientes.groupby(by=["pk_partition"])[[
"pk_cid"
]].count().sort_values(by="pk_partition", ascending=False)
df_evolucion_clientes = pd.concat([df_n_cli, _df], axis=1)
df_evolucion_clientes.columns = [
"total", "con_producto", "siguiente", "dif"
]
df_evolucion_clientes["porcentaje"] = round(
df_evolucion_clientes["con_producto"] /
df_evolucion_clientes["total"] * 100, 2)
df_evolucion_clientes.reset_index(inplace=True)
fig = px.line(df_evolucion_clientes,
x="pk_partition",
y="porcentaje",
range_y=[50, 100],
title='Evolución de la tasa de conversión')
return fig, df_evolucion_clientes
def update_file_graph(query):
try:
project_id = get_project_id(query)
with dashapp.server.app_context():
project_files = AudioFile.query.join(
AudioFile.monitoring_station
).filter(
MonitoringStation.project_id == project_id).with_entities(
AudioFile.size, AudioFile.timestamp,
MonitoringStation.name.label('station')).subquery()
bytes_per_day = db.session.query(
func.sum(project_files.c.size).label('bytes'),
func.date(project_files.c.timestamp).label('date'),
project_files.c.station).group_by('station').group_by(
'date').order_by('date')
df = pd.read_sql(bytes_per_day.statement, db.session.bind)
df['MB'] = df['bytes'] / (1024 * 1024)
station_colors = get_station_colors(project_id)
fig = px.line(df,
x='date',
y='MB',
color='station',
color_discrete_map=station_colors,
title='Data recorded per day')
fig.update_traces(mode='markers')
fig.update_layout(hovermode='x')
except:
fig = go.Figure()
fig.update_layout(title='Data recorded per day',
annotations=[{
'text': 'Loading...',
'showarrow': False,
'font': {
'size': 28
}
}])
return fig
def graph():
location = input('please Enter the location of the CVS file\n')
Xvalue = input('\nPlease enter the X Value\n')
Yvalue = input('\nPlease enter the Y Value\n')
title = input('\nPlease enter the Title of the graph\n')
color = input("\nWhat should be the color of the graph base on?\n")
graphValue = input('\nwhat type of graph Do you want?\n')
df = pd.read_csv(location)
print(df)
graph = graphValue.lower()
if 'line' in graph:
graph = pe.line(df, x=Xvalue, y=Yvalue, color=color, title=title)
graph.show()
elif 'line' in graph:
graph = pe.bar(df, x=Xvalue, y=Yvalue, title=title, color=color)
graph.show()
elif 'scatter' in graph:
graph = pe.scatter(df, x=Xvalue, y=Yvalue, color=color, title=title)
graph.show()
def main():
menu=['主页','排名进度','合并查询','语义分析','物品识别','关于']
choice = st.sidebar.selectbox('Menu',menu)
engine = create_engine('mysql+pymysql://chencheng:iKWz@4*[email protected]:3306')
catelist=pd.read_csv('Files/category__.csv')
@st.cache
def load_data(category_id):
data_sql=f'''
select category_id,category_name,asin,ranking,snapshotted_at as date
from mws_data.project_20_listings where category_id='{category_id}'
'''
data_engine = create_engine('mysql+pymysql://chencheng:iKWz@4*[email protected]:3306')
data=pd.read_sql(data_sql,data_engine)
data['date']=pd.to_datetime(pd.to_datetime(data['date']).dt.date)
return data
if choice == '主页':
st.title('傲基2.0品类管理')
st.header('品类详情')
st.markdown('针对每个asin，日排名规则为：每日排名取当日排名的中位数值')
elif choice == '排名进度':
st.title('当月排名进度')
category = st.selectbox('品类名称:',catelist['category_name'])
catid=catelist[catelist['category_name']==category]['category_id'].values[0]
st.markdown(f'#### 已选中 **{category}** , 品类ID: **{catid}**')
data = None
while data is None:
data_load_state=st.text('正在加载数据')
data=load_data(catid)
table=data.groupby(['asin','date'])['ranking'].agg('median').apply(lambda x:int(x))
t1=table.unstack(level=1)
# with open('app.html') as f:
# t1= f.read()
# components.html(t1,height=800)
st.write(t1)
data_load_state.success('✔️ 数据加载完成')
time.sleep(1)
data_load_state.text('')
chosed_asin = st.selectbox('Asin',data.asin.unique())
a1=pd.DataFrame(t1.loc[f'{chosed_asin}',:].dropna()).transpose()
a1_rank=a1.apply(lambda x: int(np.median(x)),axis=1)[0]
st.dataframe(a1)
#画图模块
t2=pd.DataFrame(t1.loc[f'{chosed_asin}'])
t2['date']=t2.index.get_level_values(0)
t2.reset_index(drop=True)
fig = px.line(t2, x=t2['date'], y=t2[f'{chosed_asin}'], title=f'{chosed_asin}的当月排名情况(红线为当月排名)')
fig.add_shape(type='line',
x0=t2['date'].iloc[0],
y0=a1_rank,
x1=t2['date'].iloc[-1],
y1=a1_rank,
line=dict(color='Red'),
xref='x',
yref='y')
st.plotly_chart(fig, use_container_width=True)
#画图模块
st.info(f'ASIN {chosed_asin}的当月排名为: {int(a1_rank)}')
a2=t1.apply(lambda x: np.median(x),axis=1).sort_values()
if int(a2[0])==1:
st.success(f'当前品类下月排名最高的Asin为{a2.index[0]}, 月排名为: {int(a2[0])}')
elif int(a2[0])<=8:
st.info(f'当前品类下月排名最高的Asin为{a2.index[0]}, 月排名为: {int(a2[0])}')
else:
st.warning(f'当前品类下月排名最高的Asin为{a2.index[0]}, 月排名为: {int(a2[0])}')
d1,d2=st.beta_columns(2)
with d1:
start_time = st.date_input("根据日期查看",t2['date'].iloc[0])
with d2:
end_time = st.date_input("结束日期",t2['date'].iloc[-1])
if str(start_time) in t1.columns:
st.write(t1.loc[:,start_time:end_time])
t2=t1[f'{start_time}'].sort_values()
top1=t2[0]
topn=t2.where(t2<=8).count()
if top1==1:
st.success(f'{start_time}日 {t2.index[0]} 的日排名为第一！')
elif topn>=4:
st.success('有四个个以上asin进入前八！')
else:
st.warning('没有排名靠前的Asin')
else:
st.write('数据缺失')
with st.beta_expander('原始数据详情',expanded=False):
i=st.number_input('输入你想要看到的条数',min_value=1,value=50,step=50)
detail=data.iloc[:i,:]
st.write(detail)
elif choice == '合并查询':
ci=st.multiselect('品类ID',data['category_id'].unique())
newdate=st.multiselect('日期',data[(data['category_id'].isin(ci))].date.unique())
asin=st.multiselect('ASIN',data[data['category_id'].isin(ci)&(data['date'].isin(newdate))].asin.unique())
newtable=data[(data['category_id'].isin(ci))&(data['asin'].isin(asin))&(data['date'].isin(newdate))]
st.write(newtable[['category_id','category_name','asin','ranking','date']])
elif choice == '语义分析':
st.subheader('亚马逊商品评论智能分析')
st.set_option('deprecation.showPyplotGlobalUse', False)
engine = create_engine('mysql+pymysql://root:aukey@[email protected]:3306')
sql1='''select distinct asin from amazon.asin_review
'''
asins = pd.read_sql(sql1,engine)
chosedasin=st.selectbox('Asin:',asins['asin'].unique())
sql=f'''
select review_star_rating,review_info, right(review_date,17)
as review_date from amazon.asin_review
where asin='{chosedasin}'
'''
review = pd.read_sql(sql,engine)
st.write(review)
no=st.number_input('输入你想要看到的评论序号',min_value=1,value=1,step=1)
ct=review.iloc[no,1]
raw_text =st.text_area('选择序号自动引用评论，或在下框手动输入文本',f'{ct}')
docx = nlp(raw_text)
c_tokens = [ token.text for token in docx ]
c_lemma = [token.lemma_ for token in docx]
c_pos = [word.tag_ for word in docx]
new_df = pd.DataFrame(zip(c_tokens,c_lemma,c_pos),columns=['Tokens','Lemma','POS'])
adj=new_df[(new_df['POS']=='JJ')|(new_df['POS']=='NN')]['Lemma'].value_counts()
wc=str(adj).replace('dtype','').replace('Lemma','').replace('int64','').replace('Name','')
c1,c2,c3,c4=st.beta_columns(4)
with c1:
qf=st.button('开始切分')
with c2:
cy=st.button('生成词云')
with c3:
fx=st.button('情感分析')
with c4:
ck=st.button('查看动名词')
if fx:
blob = TextBlob(raw_text)
sent_res=[]
sent_res.append(blob.polarity)
sent_res.append(blob.subjectivity)
st.success(f'该评论的偏向性为:{sent_res[0]},主观性为:{sent_res[1]}')
if sent_res[0]>0.3:
st.success('该客户颇为喜爱此商品！')
elif sent_res[0]>=-0.3 and sent_res[0]<=0.3:
st.info('该客户对此商品没有明显偏向')
else:
st.warning('该客户对产品很不满！')
if cy:
wordcloud = WordCloud().generate(wc)
plt.imshow(wordcloud,interpolation='bilinear')
plt.axis("off")
st.pyplot()
if ck:
st.write(pd.DataFrame(adj.rename('Count')).transpose())
if qf:
st.dataframe(new_df)
def main():
st.title('Here is Stocks')
df = pd.read_csv(
'https://robintrack.net/api/most_popular.csv?limit=100000')
df = df.head(2500).reset_index()
df['symbol_name'] = df['symbol'] + ' | ' + df['name']
# st.write(df.head(3))
# l=df['symbol_name'].to_list()
l = df['symbol'].to_list()
stocks = st.sidebar.multiselect('Select stocks', l)
# st.write(stocks)
stocks = str(stocks)
stocks = stocks.replace('[', '').replace(']',
'').replace('\'',
'').replace(' ', '')
stocks = st.sidebar.text_input(
'Or add more below (comma separated, please)', stocks)
# st.write(stocks)
# st.write(type(stocks))
# stocks = stocks.str.replace('[','').replace(']','').replace('\'','').replace(' ','')
# st.write(stocks)
stocks = list(stocks.split(','))
# st.write(type(stocks))
# st.write(stocks)
# st.write(stocks)
# s=df.loc[df.symbol_name.isin(stocks)]
# s=s['symbol'].to_list()
# more_stocks = st.sidebar.text_input('Add more below',str(s))
# more_stocks = [more_stocks]
# st.write(more_stocks)
# st.write([more_stocks])
import datetime
end = datetime.date.today()
start = datetime.date.today() - datetime.timedelta(days=30)
start_date = st.sidebar.date_input('Start date', start)
end_date = st.sidebar.date_input('End date', end)
# if start_date < end_date:
# st.success('Start date: `%s`\n\nEnd date:`%s`' % (start_date, end_date))
# else:
# st.error('Error: End date must fall after start date.')
# start = '2020-07-01'
# end = '2020-07-21'
if st.sidebar.button('Go'):
if len(stocks) > 0:
dft = dr.get_data_yahoo(stocks, start_date, end_date)
dft = dft.stack()
dft = dft.reset_index()
dft = dft.groupby('Symbols').apply(pct_change)
# st.write(dft)
fig = px.line(dft, x='Date', y='Close', color='Symbols')
# fig.show()
st.plotly_chart(fig)
fig = px.line(dft, x='Date', y='Pct Chg', color='Symbols')
st.plotly_chart(fig)
x = dft.groupby('Date').agg({'Close': 'sum'}).reset_index()
x['portfolio_pct_chg'] = 100 * (1 - x.iloc[0].Close / x.Close)
dft = pd.merge(dft,
x,
how='inner',
left_on='Date',
right_on='Date')
fig = px.line(dft, x='Date', y='portfolio_pct_chg')
st.plotly_chart(fig)
st.write(dft)
color=color_value)
else:
plot = px.scatter(data_frame=df, x=x_value, y=y_value)
# display chart in streamlit
st.plotly_chart(plot)
if chart_select == 'Histogram':
st.sidebar.subheader("Settings for Histogram")
x = st.sidebar.selectbox(label="Feature", options=numeric_columns)
bin_size = st.sidebar.slider(label="Number of bins",
min_value=10,
max_value=100,
value=50)
plot = px.histogram(data_frame=df, x=x, nbins=bin_size)
st.plotly_chart(plot)
if chart_select == 'Lineplots':
st.sidebar.subheader("Settings for Line plots.")
x_value = st.sidebar.selectbox(label='X axis', options=numeric_columns)
y_value = st.sidebar.selectbox(label='Y axis', options=numeric_columns)
plot = px.line(data_frame=df, x=x_value, y=y_value)
# display the chart
st.plotly_chart(plot)
except Exception as e:
print(e)
