def andrews_curve(
data: pd.DataFrame,
xvar: str = "t",
yvar: str = "curve_value",
targetvar: str = "target",
samplevar: str = "sample",
w: int = 450,
h: int = 300,
) -> alt.LayerChart:
selection = alt.selection_single(fields=[targetvar], bind="legend")
base = alt.Chart(data).properties(width=w, height=h).mark_line()
background_chart = base.encode(
x=alt.X(f"{xvar}:Q",
axis=alt.Axis(title=None),
scale=alt.Scale(nice=False)),
y=alt.Y(f"{yvar}:Q", axis=alt.Axis(title=None)),
detail=alt.Detail(f"{samplevar}:N"),
color=alt.value(COLORS["light_gray"]),
)
chart = background_chart.encode(color=alt.condition(
selection,
f"{targetvar}:N",
alt.value("transparent"),
legend=alt.Legend(title=f"{targetvar.title()} (click to highlight)"),
), ).add_selection(selection)
return background_chart + chart
def parallel_coordinates(data,
class_column,
cols=None,
alpha=None,
width=450,
height=300,
interactive=True,
var_name="variable",
value_name="value",
**kwds):
"""
Parallel coordinates plotting.
Parameters
----------
frame: DataFrame
class_column: str
Column name containing class names
cols: list, optional
A list of column names to use
alpha: float, optional
The transparency of the lines
interactive : bool, optional
if True (default) then produce an interactive plot
width : int, optional
the width of the plot in pixels
height : int, optional
the height of the plot in pixels
var_name : string, optional
the legend title
value_name : string, optional
the y-axis label
Returns
-------
chart: alt.Chart object
The altair representation of the plot.
See Also
--------
pandas.plotting.parallel_coordinates : matplotlib version of this routine
"""
if kwds:
warnings.warn("Unrecognized keywords in pdvega.scatter_matrix: {0}"
"".format(list(kwds.keys())))
# Transform the dataframe to be used in Vega-Lite
if cols is not None:
data = data[list(cols) + [class_column]]
cols = data.columns
df = data.reset_index()
index = (set(df.columns) - set(cols)).pop()
assert index in df.columns
df = df.melt([index, class_column],
var_name=var_name,
value_name=value_name)
chart = alt.Chart(df).properties(width=width, height=height)
chart = chart.mark_line().encode(
x=alt.X(field=var_name, type=infer_vegalite_type(df[var_name])),
y=alt.Y(field=value_name, type=infer_vegalite_type(df[value_name])),
color=alt.Color(field=class_column,
type=infer_vegalite_type(df[class_column])),
detail=alt.Detail(field=index, type=infer_vegalite_type(df[index])))
if alpha is None and df[class_column].nunique() > 20:
alpha = 0.3
if alpha is not None:
assert 0 <= alpha <= 1
return chart.encode(opacity=alt.value(alpha))
return chart
line = pd.DataFrame({ #create fake df so line can get drawn
'line_x': [0,14000],
'line_y': [0,14000],
})
line_plot = alt.Chart(line).mark_line(
opacity=0.3).encode( #draw fake line plot
x='line_x', y='line_y')
st.altair_chart(usage_2019_vs_2020 + line_plot, use_container_width=True)
st.subheader("Bring in 1figr Tiers from 2018")
figr_Tier_hist = alt.Chart(df[filt].reset_index()).mark_bar().encode(
alt.X('1figr Tier:O'),
alt.Y('count()'),
alt.Detail('index:Q'),
tooltip=['1figr Tier', 'Title Name', 'CPU_2020'],
color=alt.Color('Decision:N',
scale=subscribed_colorscale) #Nominal data type
).interactive().properties(
height=400,
title={
"text": ["Histogram of 1figr Tiers"],
"subtitle": [
"Titles stacked by Tier",
"NaN and 0 mean Tier was not available for that title"
],
"color":
"black",
"subtitleColor":
"gray"
def andrews_curves(data,
class_column,
samples=200,
alpha=None,
width=450,
height=300,
**kwds):
"""
Generates an Andrews curves visualization for visualising clusters of
multivariate data.
Andrews curves have the functional form:
f(t) = x_1/sqrt(2) + x_2 sin(t) + x_3 cos(t) +
x_4 sin(2t) + x_5 cos(2t) + ...
Where x coefficients correspond to the values of each dimension and t is
linearly spaced between -pi and +pi. Each row of frame then corresponds to
a single curve.
Parameters:
-----------
data : DataFrame
Data to be plotted, preferably normalized to (0.0, 1.0)
class_column : string
Name of the column containing class names
samples : integer
Number of points to plot in each curve
alpha: float, optional
The transparency of the lines
width : int, optional
the width of the plot in pixels
height : int, optional
the height of the plot in pixels
**kwds: keywords
Additional options
Returns:
--------
chart: alt.Chart object
"""
if kwds:
warnings.warn("Unrecognized keywords in pdvega.andrews_curves(): {0}"
"".format(list(kwds.keys())))
t = np.linspace(-np.pi, np.pi, samples)
vals = data.drop(class_column, axis=1).values.T
curves = np.outer(vals[0], np.ones_like(t))
for i in range(1, len(vals)):
ft = ((i + 1) // 2) * t
if i % 2 == 1:
curves += np.outer(vals[i], np.sin(ft))
else:
curves += np.outer(vals[i], np.cos(ft))
df = pd.DataFrame({
"t":
np.tile(t, curves.shape[0]),
"sample":
np.repeat(np.arange(curves.shape[0]), curves.shape[1]),
" ":
curves.ravel(),
class_column:
np.repeat(data[class_column], samples),
})
chart = alt.Chart(df).properties(width=width, height=height).mark_line()
chart = chart.encode(
x=alt.X(field="t", type="quantitative"),
y=alt.Y(field=" ", type="quantitative"),
color=alt.Color(field=class_column,
type=infer_vegalite_type(df[class_column])),
detail=alt.Detail(field='sample', type="quantitative"))
if alpha is None and df[class_column].nunique() > 20:
alpha = 0.5
if alpha is not None:
assert 0 <= alpha <= 1
return chart.encode(opacity=alt.value(alpha))
return chart
timeUnit='year',
scale=alt.Scale(range=['green', 'yellow', 'blue']))).facet(
'primary', columns=2)
chart
"""## <b>V2</b>
### Graph with details and tooltips
#### <b>Pros</b>
All the necessary data is present
#### <b>Cons</b>
Too much data, hard to follow and understand
"""
chart = base.mark_circle(size=80).encode(y=alt.Y('average(score)'),
x=alt.X('count()'),
detail=alt.Detail('date:T',
timeUnit='year'),
tooltip=alt.Tooltip('date:T',
timeUnit='year'),
color=alt.Color('primary'))
chart
"""## <b>V3</b>
### Interactive graphs - click on upper's graph genre to get data by year in the lower graph (both graphs have tooltips)
#### <b>Pros</b>
Flexibility, not too much data
#### <b>Cons</b>
Takes time to get used to, hard to compare particular years
"""
selector = alt.selection_single(fields=['primary'])
def _alt_detail(self):
return alt.Detail(f'{self._detailby}:N')
"High Cost-per-Use rank (least economical) journals show up in darker colors"
],
"color":
"black",
"subtitleColor":
"gray"
}).add_selection(selection2)
st.altair_chart(weighted_vs_cost2, use_container_width=True)
st.subheader(
'Look into the authorships, citations, and downloads of each journal')
auth_hist = alt.Chart(df[filt].reset_index()).mark_bar(width=10).encode(
alt.X('authorships:Q',
title="Authorships (average per year over the next five years)"),
alt.Y('count()', axis=alt.Axis(grid=False)),
alt.Detail('index'),
tooltip=['title', 'authorships', 'subscription_cost', 'subscribed'],
color=alt.Color('subscribed:N', scale=subscribed_colorscale)
).interactive().properties(
height=400,
title={
"text": ["Authorships Distribution"],
"subtitle": [
"What do the range of Authorships look like?",
"Use this graph to help set the Authorships slider filter and narrow down titles of interest"
],
"color":
"black",
"subtitleColor":
"gray"
})
############################
#line weekday chart
############################
source = df_per_weekday
df_per_weekday['option_chosen']='other species'
df_per_weekday.loc[df_per_weekday['species']==option, 'option_chosen'] = option
line=alt.Chart(source).mark_line(point=True).encode(
x='weekday_char',
y=alt.Y('dev_pct_from_avg_price', axis=alt.Axis(format='+%')),
detail=alt.Detail('species'),
tooltip='species',
color=alt.Color('option_chosen', scale=alt.Scale(domain=['other species', option]
, range = ['#cfebfd','#0C266A']), legend =None)
).interactive().properties(
height=450
)
# layer that accomplishes the highlighting
source_highlight = df_per_weekday[df_per_weekday["species"] == option]
line_highlight = alt.Chart(source_highlight).mark_line(point=True, size =3).encode(
x=alt.X('weekday_char', title = 'Dia de la Semana'),
y=alt.Y('dev_pct_from_avg_price', title = 'Desviacion del precio medio'),
detail=alt.Detail('species'),
tooltip='species',
color=alt.Color('option_chosen', scale=alt.Scale(domain=['other species', option]