def plot_facet_scatter(df, x, y, group1, group2):
fig = ff.create_facet_grid(
df,
x=x,
y=y,
color_name=group2,
color_is_cat=True if df[group2].dtype.name == 'category' else False,
# facet_col=group1,
facet_row=group1,
)
plotly.offline.plot(fig)
# plot_box(df, 'Species', 'sepal length (cm)')
# plot_scatter(df, 'sepal length (cm)', 'sepal width (cm)', 'Species')
# plot_scatter(df, 'petal length (cm)', 'sepal width (cm)', 'Species')
# plot_density(df, 'petal width (cm)', 'Species')
#
# boston = load_boston()
# df1 = pd.DataFrame(boston.data, columns=boston.feature_names)
# df1['Target'] = boston.target
# df1['RAD'] = df1['RAD'].astype('category')
# plot_box(df1, 'RAD', 'Target')
#
# ###PAIR PLOTS
# sns.pairplot(df1)
# plt.show()
#
#
# plot_facet_scatter(df1, 'AGE', 'Target', 'RAD', 'ZN')
def update_facet_grid1(traceType, y):
facet_hist = ff.create_facet_grid(
df,
y=y,
facet_row='Gender',
facet_col='Capture Location',
trace_type=traceType,
)
facet_hist['layout']['title'] = '{} - {} by gender and location'.format(
traceType, y)
return dcc.Graph(
id='facet_grid_1',
figure=facet_hist,
)
def redrawGraph(x, y, color, row, col, size, checklist, prevLayout):
df = deepcopy(diamonds[:size])
df = df.reset_index(drop=True)
rowVal = row
colVal = col
trace = 'scattergl'
if (row == 'None'):
rowVal = None
if (col == 'None'):
colVal = None
if (color == 'None'):
color = None
if ('jitter' in checklist):
df = pd.concat([deepcopy(df), jitter(df, x, y)], ignore_index=True)
fig = ff.create_facet_grid(df,
x=x,
y=y,
trace_type='scattergl',
color_name=color,
facet_row=rowVal,
facet_col=colVal,
marker=dict(
size=4,
line=dict(width=0.1, color='black'),
),
ggplot2=True)
fig['layout']['hovermode'] = 'closest'
fig['layout']['dragmode'] = 'zoom'
fig['layout']['height'] = 900
fig['layout']['width'] = None
if not rowVal and not colVal:
fig['layout']['yaxis']['zeroline'] = False
fig['layout']['xaxis']['zeroline'] = False
fig['layout']['yaxis']['title'] = y
fig['layout']['annotations'][1]['text'] = ""
fig['layout']['xaxis']['title'] = x
fig['layout']['annotations'][0]['text'] = ""
fig['layout']['autosize'] = True
if discrete(x):
relabel(x, fig, 0)
if discrete(y):
relabel(y, fig, 1)
print(fig)
return fig
def q1():
D = {
(1, 1): 'chest_pain_type',
(1, 2): 'resting_blood_pressure',
(1, 3): 'cholesterol',
(1, 4): 'fasting_blood_sugar',
(2, 1): 'rest_ecg',
(2, 2): 'max_heart_rate_achieved',
(2, 3): 'exercise_induced_angina',
(3, 1): 'st_depression',
(3, 2): 'st_slope',
(3, 3): 'num_major_vessels',
(4, 1): 'thalassemia'
}
# q1_fig = tools.make_subplots(rows=4, cols=4)
q1_fig = plotly.tools.make_subplots(rows=4,
cols=4,
shared_xaxes=True,
subplot_titles=[
'CPT', 'RBP', D[(1, 3)], 'FBS',
'RE', 'MHRA', 'EIA', '', D[(3, 1)],
D[(3, 1)], 'NMV', '', D[(4, 1)],
'', '', ''
])
q1_df = df.astype('float64')
q1_df['sex'] = q1_df['sex'].apply(lambda x: 'Male'
if x == 1.0 else 'Female')
# for k in q1_df.keys():
for k in D.keys():
trace = ff.create_facet_grid(
q1_df,
x='age',
y=D[k],
color_name='sex',
# show_boxes=False,
marker={
'size': 5,
'opacity': 1.0
},
colormap={
'Male': 'rgb(165, 242, 242)',
'Female': 'rgb(253, 174, 216)'
},
facet_col_labels='name',
facet_row_labels='name',
# ggplot2=True
)
# print(k[0], k[1])
for f in trace.data:
q1_fig.append_trace(f, k[0], k[1])
# fig = ff.create_facet_grid(
# q1_df,
# x='age',
# y='thalassemia',
# color_name='sex',
# show_boxes=False,
# marker={'size': 10, 'opacity': 1.0},
# colormap={'Male': 'rgb(165, 242, 242)', 'Female': 'rgb(253, 174, 216)'}
# )
# py.iplot(fig, filename='facet - custom colormap')
q1_fig['layout'].update(
height=700,
width=700,
showlegend=False,
title='<b>Statistics of Attributes<b>',
# title="<b>Feature importance</b>",
titlefont=dict(size=20, color='rgb(23,203,203)'),
)
return q1_fig
def create_facet_grid(*args, **kwargs):
FigureFactory._deprecated('create_facet_grid')
from plotly.figure_factory import create_facet_grid
return create_facet_grid(*args, **kwargs)
def redrawGraph(x, y, color, row, col, size, checklist, prevLayout):
df = deepcopy(diamonds[:size])
df = df.reset_index(drop=True)
rowVal = row
colVal = col
colorVal = color
trace = 'scattergl'
if (row == 'None'):
rowVal = None
if (col == 'None'):
colVal = None
if (color == 'None'):
colorVal = None
if ('jitter' in checklist):
df = pd.concat([deepcopy(df), jitter(df, x, y)], ignore_index=True)
fig = ff.create_facet_grid(df,
x=x,
y=y,
trace_type='scattergl',
color_name=colorVal,
facet_row=rowVal,
facet_col=colVal,
marker=dict(
size=4,
line=dict(width=0.1, color='black'),
),
ggplot2=True)
fig['layout']['hovermode'] = 'closest'
fig['layout']['dragmode'] = 'zoom'
fig['layout']['height'] = 900
fig['layout']['width'] = None
if not rowVal and not colVal:
fig['layout']['yaxis']['zeroline'] = False
fig['layout']['xaxis']['zeroline'] = False
fig['layout']['yaxis']['title'] = y
fig['layout']['annotations'][1]['text'] = ""
fig['layout']['xaxis']['title'] = x
fig['layout']['annotations'][0]['text'] = ""
fig['layout']['autosize'] = True
# hack solution to legendgroup'n
if colorVal is not None:
if (colorVal != rowVal and colorVal != colVal):
for j in range(0, len(fig['data'])):
fig['data'][j]['legendgroup'] = fig['data'][j]['name']
fig['data'][j]['showlegend'] = False
if (color == 'color-labels'):
for k in range(0, len(df['color-labels'].unique())):
fig['data'][k]['showlegend'] = True
elif (color == 'cut-labels'):
for k in range(0, len(df['cut-labels'].unique())):
fig['data'][k]['showlegend'] = True
elif (color == 'clarity-labels'):
for k in range(0, len(df['clarity-labels'].unique())):
fig['data'][k]['showlegend'] = True
elif rowVal is not None and colVal is not None and rowVal != colVal and\
(colorVal == rowVal or colorVal == colVal):
for j in range(0, len(fig['data'])):
fig['data'][j]['legendgroup'] = fig['data'][j]['name']
fig['data'][j]['showlegend'] = False
if (color == 'color-labels'):
l = 0
for k in range(0, len(df['color-labels'].unique())):
fig['data'][l]['showlegend'] = True
l = l + (len(df['color-labels'].unique()) + 1)
elif (color == 'cut-labels'):
l = 0
for k in range(0, len(df['cut-labels'].unique())):
fig['data'][l]['showlegend'] = True
l = l + (len(df['cut-labels'].unique()) + 1)
elif (color == 'clarity-labels'):
l = 0
for k in range(0, len(df['clarity-labels'].unique())):
fig['data'][l]['showlegend'] = True
l = l + (len(df['clarity-labels'].unique()) + 1)
if discrete(x):
relabel(x, fig, 0)
if discrete(y):
relabel(y, fig, 1)
return fig
def make_subplots(data: pd.DataFrame,
columns: List[str] = None,
*,
kind: str = "box",
**kwargs):
"""Make subplots and arrange them in an optimized grid layout."""
if kind not in ("box", "histogram", "scatter", "scatter_with_bounds"):
raise ValueError(f"Can NOT handle plot of kind: {kind}.")
index = data.index.droplevel(-1).unique()
if len(index.names) > 2:
logger.warning(
f"Can only handle hierarchical index of depth <= 2, got {len(index.names)}. Grouping index."
)
return make_subplots(group_index(data, range(index.nlevels - 1)),
columns,
kind=kind,
**kwargs)
grid = ff.create_facet_grid(
data.reset_index(),
facet_row=index.names[1] if index.nlevels > 1 else None,
facet_col=index.names[0],
trace_type="box", # box does not need data specification
ggplot2=True,
)
shape = np.shape(grid._grid_ref)[:-1]
sub_plots = tools.make_subplots(
rows=shape[0],
cols=shape[1],
shared_yaxes=kwargs.pop("shared_yaxes", True),
shared_xaxes=kwargs.pop("shared_xaxes", False),
print_grid=kwargs.pop("print_grid", False),
)
if isinstance(index, pd.MultiIndex):
index_grid = zip(*index.labels)
else:
index_grid = iter(
np.transpose([
np.tile(np.arange(shape[1]), shape[0]),
np.repeat(np.arange(shape[0]), shape[1])
]))
for idx, grp in data.groupby(level=np.arange(index.nlevels).tolist()):
if not isinstance(columns, str) and kind == "scatter_with_bounds":
if columns is None:
raise ValueError(
"`scatter_with_bounds` requires `col` argument, not provided."
)
try:
columns, = columns
except ValueError:
raise ValueError(
"`scatter_with_bounds` does not allow for multiple columns."
)
fig = eval(f"create_duration_{kind}(grp, columns, **kwargs)")
col, row = map(int, next(index_grid)) # col-first plotting
for trace in fig.data:
sub_plots.append_trace(trace, row + 1, col + 1)
layout = sub_plots.layout
layout.update(
title=kwargs.get("title", fig.layout.title),
shapes=grid.layout.shapes,
annotations=grid.layout.annotations,
showlegend=False,
)
x_dom_vals = [k for k in layout.to_plotly_json().keys() if "xaxis" in k]
y_dom_vals = [k for k in layout.to_plotly_json().keys() if "yaxis" in k]
layout_shapes = pd.DataFrame(
layout.to_plotly_json()["shapes"]).sort_values(["x0", "y0"])
h_shapes = layout_shapes[~layout_shapes.x0.duplicated(keep=False)]
v_shapes = layout_shapes[~layout_shapes.y0.duplicated(keep=False)]
# handle single-columns
h_shapes = h_shapes.query("y1 - y0 != 1")
v_shapes = v_shapes.query("x1 - x0 != 1")
# update axis domains and layout
for idx, x_axis in enumerate(x_dom_vals):
x0, x1 = h_shapes.iloc[idx % shape[1]][["x0", "x1"]]
layout[x_axis].domain = (x0 + 0.03, x1 - 0.03)
layout[x_axis].update(showticklabels=False, zeroline=False)
for idx, y_axis in enumerate(y_dom_vals):
y0, y1 = v_shapes.iloc[idx % shape[0]][["y0", "y1"]]
layout[y_axis].domain = (y0 + 0.03, y1 - 0.03)
layout[y_axis].update(zeroline=False)
# correct annotation to match the relevant group and width
annot_df = pd.DataFrame(
layout.to_plotly_json()["annotations"]).sort_values(["x", "y"])
annot_df = annot_df[annot_df.text.str.len() > 0]
aw = min( # annotation width magic
int(max(60 / shape[1] - (2 * shape[1]), 6)),
int(max(30 / shape[0] - (2 * shape[0]), 6)))
for i, annot_idx in enumerate(annot_df.index):
annot = layout.annotations[annot_idx]
index_label: Union[str, Any] = annot["text"]
if isinstance(index, pd.MultiIndex):
index_axis = i >= shape[1]
if shape[0] == 1:
pass # no worries, the order and label are aight
elif shape[1] == 1:
index_label = index.levels[index_axis][max(0, i - 1)]
else:
index_label = index.levels[index_axis][i % shape[1]]
text: str = str(index_label)
annot["text"] = re.sub(r"^(.{%d}).*(.{%d})$" % (aw, aw),
"\g<1>...\g<2>",
text) # Ignore PycodestyleBear (W605)
annot["hovertext"] = "<br>".join(pformat(index_label).split("\n"))
# add axis titles as plot annotations
layout.annotations = (
*layout.annotations,
{
"x": 0.5,
"y": -0.05,
"xref": "paper",
"yref": "paper",
"text": fig.layout.xaxis["title"]["text"],
"showarrow": False,
},
{
"x": -0.05,
"y": 0.5,
"xref": "paper",
"yref": "paper",
"text": fig.layout.yaxis["title"]["text"],
"textangle": -90,
"showarrow": False,
},
)
# custom user layout updates
user_layout = kwargs.pop("layout", None)
if user_layout:
layout.update(user_layout)
return sub_plots
def update_graph(table, xaxis_field, yaxis_fields, split, chart_type,
aggregation_type, split_type, filter_variable_1,
filter_values_1, filter_variable_2, filter_values_2,
filter_variable_3, filter_values_3):
if xaxis_field is None or yaxis_fields is None:
return {"layout": {'style': {'display': 'none'}}}
filter_variables = [
filter_variable_1, filter_variable_2, filter_variable_3
]
filter_values = [filter_values_1, filter_values_2, filter_values_3]
filters = {}
for i, filter_variable in enumerate(filter_variables):
if filter_variable is not None and filter_values[i] is not None:
filters[filter_variable] = filter_values[i]
if split is None:
columns = set(yaxis_fields + [xaxis_field] + list(filters.keys()))
df = pd.read_sql(
'SELECT ' + ','.join(list(columns)) + ' FROM ' + table, conn)
for variable, values in filters.items():
df = df[df[variable].isin(values)]
if aggregation_type == 'avg':
df = df.groupby('step', as_index=False).aggregate(np.mean)
elif aggregation_type == 'min':
df = df.groupby('step', as_index=False).aggregate(min)
elif aggregation_type == 'max':
df = df.groupby('step', as_index=False).aggregate(max)
elif aggregation_type == 'median':
df = df.groupby('step', as_index=False).aggregate(np.median)
elif aggregation_type == 'count':
df = df.groupby('step', as_index=False).aggregate(len)
if chart_type == 'bar':
plot = [
go.Bar(x=df[xaxis_field], y=df[yaxis_field], name=yaxis_field)
for yaxis_field in yaxis_fields
]
elif chart_type == 'box':
plot = [
go.Box(x=df[xaxis_field], y=df[yaxis_field], name=yaxis_field)
for yaxis_field in yaxis_fields
]
elif chart_type == 'histogram':
plot = [
go.Histogram(y=df[yaxis_field],
name=yaxis_field,
histnorm='probability')
for yaxis_field in yaxis_fields
]
else:
plot = [
go.Scatter(x=df[xaxis_field],
y=df[yaxis_field],
mode=chart_type,
name=yaxis_field,
marker={
'size': 15,
'opacity': 0.5,
'line': {
'width': 0.5,
'color': 'white'
}
}) for yaxis_field in yaxis_fields
]
else:
columns = set(yaxis_fields + [xaxis_field] + split +
list(filters.keys()))
df = pd.read_sql(
'SELECT ' + ','.join(list(columns)) + ' FROM ' + table, conn)
for variable, values in filters.items():
df = df[df[variable].isin(values)]
if aggregation_type == 'avg':
df = df.groupby(split + ['step'],
as_index=False).aggregate(np.mean)
elif aggregation_type == 'min':
df = df.groupby(split + ['step'], as_index=False).aggregate(min)
elif aggregation_type == 'max':
df = df.groupby(split + ['step'], as_index=False).aggregate(max)
elif aggregation_type == 'median':
df = df.groupby(split + ['step'],
as_index=False).aggregate(np.median)
elif aggregation_type == 'count':
df = df.groupby(split + ['step'], as_index=False).aggregate(len)
if split_type == 'facet':
trace = 'scatter' if chart_type == 'markers' or chart_type == 'lines' else chart_type
if len(split) == 1:
plot = ff.create_facet_grid(
df,
x=xaxis_field,
y=yaxis_fields[0],
facet_col=split[0],
trace_type=trace,
)
elif len(split) == 2:
plot = ff.create_facet_grid(
df,
x=xaxis_field,
y=yaxis_fields[0],
facet_col=split[0],
facet_row=split[1],
trace_type=trace,
)
else:
plot = ff.create_facet_grid(
df,
x=xaxis_field,
y=yaxis_fields[0],
facet_col=split[0],
facet_row=split[1],
color_name=split[2],
color_is_cat=True,
trace_type=trace,
)
else:
levels = df[split[0]].unique()
if chart_type == 'bar':
plot = [
go.Bar(x=df[df[split[0]] == level][xaxis_field],
y=df[df[split[0]] == level][yaxis_fields[0]],
name=level) for level in levels
]
elif chart_type == 'box':
plot = [
go.Box(x=df[df[split[0]] == level][xaxis_field],
y=df[df[split[0]] == level][yaxis_fields[0]],
name=level) for level in levels
]
elif chart_type == 'histogram':
plot = [
go.Histogram(y=df[df[split[0]] == level][yaxis_fields[0]],
name=level,
histnorm='probability') for level in levels
]
else:
plot = [
go.Scatter(x=df[df[split[0]] == level][xaxis_field],
y=df[df[split[0]] == level][yaxis_fields[0]],
mode=chart_type,
name=level,
marker={
'size': 15,
'opacity': 0.5,
'line': {
'width': 0.5,
'color': 'white'
}
}) for level in levels
]
layout = {'style': {'display': 'block'}}
return {'data': plot, 'layout': layout}
def create_facet_grid(*args, **kwargs):
FigureFactory._deprecated('create_facet_grid')
from plotly.figure_factory import create_facet_grid
return create_facet_grid(*args, **kwargs)
fig['layout'].update(showlegend=False,
title="Price Distributions by Marital Status",
height=1000,
width=800)
iplot(fig, filename='custom-sized-subplot-with-subplot-titles')
df.head()
# Notice how divorced have a considerably low amount of balance.
fig = ff.create_facet_grid(df,
x='duration',
y='balance',
color_name='marital',
show_boxes=False,
marker={
'size': 10,
'opacity': 1.0
},
colormap={
'single': 'rgb(165, 242, 242)',
'married': 'rgb(253, 174, 216)',
'divorced': 'rgba(201, 109, 59, 0.82)'
})
iplot(fig, filename='facet - custom colormap')
# Hmmm We have missed some important clients with some high balances.
# This shouldn't be happening.
fig = ff.create_facet_grid(
df,
y='balance',
facet_row='marital',
facet_col='deposit',
linewidth=1,
palette="Set2")
ax3.legend_.remove()
ax3.set_title("Relationship between Smokers and Charges")
plt.show()
# Make sure we don't have any null values
df[df.isnull().any(axis=1)]
fig = ff.create_facet_grid(df,
x='age',
y='charges',
color_name='weight_condition',
show_boxes=False,
marker={
'size': 10,
'opacity': 1.0
},
colormap={
'Underweight': 'rgb(208, 246, 130)',
'Normal Weight': 'rgb(166, 246, 130)',
'Overweight': 'rgb(251, 232, 238)',
'Obese': 'rgb(253, 45, 28)'
})
251, 232, 238
fig['layout'].update(title="Weight Status vs Charges",
width=800,
height=600,
plot_bgcolor='rgb(251, 251, 251)',
paper_bgcolor='rgb(255, 255, 255)')
def update_figure(n_clicks, episode_slider, word_input, speaker_input):
fullmerge = {}
data_subset = data[data['episode'] >= episode_slider[0]][data['episode'] <= episode_slider[1]]
#Merge all episodes into one long string
peeps = list(filter(None,speaker_input.upper().replace(" ","").split(",")))
for peep in peeps:
fullmerge.update({peep:' '.join(data_subset[data_subset.Speaker == peep]['cleaned'])})
finaldata = pd.DataFrame.from_dict(fullmerge, orient = 'index').reset_index().rename(columns = {'index':'speaker',0:'text'})
totalwords = pd.DataFrame()
for peep in peeps:
totalwords = totalwords.append(pd.Series([peep, ' ', finaldata[finaldata['speaker'] == peep].text.str.count(' ').iloc[0]]), ignore_index = True)
totalwords = totalwords.rename(index=str, columns={0: "speaker", 1: "word",2: "total"})
words = list(filter(None,word_input.lower().split(",")))
#Calculate frequency of words in finaldata
df = pd.DataFrame()
for peep in peeps:
for word in words:
df = df.append(pd.Series([peep, word, finaldata[finaldata['speaker'] == peep].text.str.count(word).iloc[0]]), ignore_index = True)
df = df.rename(index=str, columns={0: "speaker", 1: "word",2: "amount"})
#Calculate rate per 1000 words
df = pd.merge(df, totalwords[['speaker','total']], on='speaker')
df['Number of times said per 1000 words'] = (df['amount']/df['total'])*1000
#Sort data by rate and then speaker
df = df.sort_values(by=['Number of times said per 1000 words','speaker'], ascending = [True,False])
#Graph
if len(words) != 1:
if len(peeps) != 1:
#Apply the tfidf function, and find the most "distinguishing" word among the given words and speakers
tfidf = TfidfVectorizer(stop_words='english', vocabulary = words)
tfs = tfidf.fit_transform(finaldata['text'])
matrix = pd.DataFrame(tfs.todense(), index = peeps, columns = tfidf.get_feature_names()).transpose()
matrix['word'] = matrix.index
matrix = pd.melt(matrix, id_vars = 'word')
matrix = matrix.rename(index=str, columns={'value': "tfidf",'variable': "speaker"})
distWord = matrix.loc[matrix['tfidf'].idxmax()]['word']
distSpeaker = matrix.loc[matrix['tfidf'].idxmax()]['speaker']
tfidfSent = ("Most distinguishing: '" + distWord + "' by " + distSpeaker + ".*")
fig = ff.create_facet_grid(
df,
x='Number of times said per 1000 words',
y='word',
facet_col='speaker',
color_name='speaker',
trace_type='bar',
orientation = 'h',
scales = 'free',
width = 1200
)
for i in range(len(peeps)+1):
if i == 0:
fig.layout.xaxis.update({'range': [df['Number of times said per 1000 words'].min(), (df['Number of times said per 1000 words'].max()+(.15 * df['Number of times said per 1000 words'].max()))]})
else:
exec('fig.layout.xaxis' + str(i)+".update({'range': [df['Number of times said per 1000 words'].min(), (df['Number of times said per 1000 words'].max()+(.15 * df['Number of times said per 1000 words'].max()))]})")
fig.layout.xaxis.title = tfidfSent
fig.layout.update(plot_bgcolor='rgba(230,230,230,90)')
elif len(peeps) == 1:
fig = ff.create_facet_grid(
df,
x='word',
y='Number of times said per 1000 words',
color_name='word',
trace_type='bar',
scales = 'free',
width = 1200
)
fig.layout.update(plot_bgcolor='rgba(230,230,230,90)')
elif len(words) == 1:
fig = ff.create_facet_grid(
df,
x='speaker',
y='Number of times said per 1000 words',
color_name='speaker',
trace_type='bar',
scales = 'free',
width = 1200
)
fig.layout.update(plot_bgcolor='rgba(230,230,230,90)')
return {
'data': fig
}
def faceted_fig(df, x, y, facet_col, y_bounds=None):
fig = ff.create_facet_grid(df, x=x, y=y,
facet_col=facet_col,
facet_col_labels='name',
facet_row_labels='name')
return fig
