def waterfall(s,
color_coded=False,
annotate=None,
up_name='positive',
down_name='negative'):
"""
Firxt category axis automatically not float.
Parameters
----------
s: pd.Series
color_coded: boolean, default to False
Whether to color coded negative and positive values
annotate: {'inside', 'outside', None}, default to None
annotate the bar with actual value.
up_name,down_name: string
increase/decrease bar name. Only work if color_coded is True
Returns
-------
Chart Object
"""
if annotate not in [None, 'inside', 'outside']:
raise ValueError("invalid parameters")
return make_chart(s,
type='waterfall',
up_name=up_name,
down_name=down_name,
color_coded=color_coded,
annotate=annotate)
def scatter(df, x, y, s=None, c=None, saturate=None, size_px=(10, 70)):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x,y: string, columns in pd.DataFrame
Used as coordinate in scatter chart
s: string, columns in pd.DataFrame default None
Used as sizing value of the scatter points
c: string, default None
column used as grouping color category
saturate: string, default None
column to use for saturation
size_px: tuple, default (10,70)
boundary size, lower and upper limit in pixel for min-max scatter points
Returns
-------
Chart Object
"""
return make_chart(df,
type='scatter',
x=x,
y=y,
s=s,
c=c,
saturate=saturate,
size_px=size_px)
def bar_tidy(df, full=False, stacked=False, trendline=False, annotate=None):
"""
This plot assume DataFrame can be directly consumed (tidy data). Used for
customized manipulation data that normal plot can't provides.
* data is only 2-dimension. There is no hierarchical columns or index.
* data is cleaned and aggregated.
* no duplicate values for each index-column pair (tidy)
* index is used for category x-axis.
* each column corresponds to one series.
* values in one column belong to column data series.
df: pd.DataFrame
data to be used for the chart
stacked: boolean, default False.
Whether to stacked category on top of the other categories.
full: boolean, default False.
If true, set to full area stacked chart. Only work if stacked is True.
trendline: boolean, default False.
If true, add line that connected the bars. Only work if not category,
category but stacked, or not full.
annotate: string, {'all', True, None} default None
if True, annotate value on top of the plot element. If stacked is
also True, annotate the last category. if 'all' and stacked,
annotate all category
Returns
-------
Chart Object
"""
return make_chart(df,
x=df.index.name,
c="unnamed",
stacked=stacked,
full=full,
type='bar_tidy',
trendline=trendline,
annotate='top' if annotate is True else annotate)
def line(df,
x,
y=None,
c=None,
how=None,
stacked=False,
area=False,
annotate=None,
full=False,
smooth=False,
sort_on='index',
sort_c_on=None,
ascending=True):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x: string
columns to be used as category axis
y: string, default None
if None, use count of category. otherwise aggregate based on y columns
c: string, default None
category column inside x-axis
how: string, default None
to be passed to pd.group_by(x).aggregate(how). Can be mean,median,
or any reduced operations.
stacked: boolean, default False.
Whether to stacked category on top of the other categories.
area: boolean, default False.
Whether to fill the area with line colors.
annotate: string, {'all',True} default None
if True, annotate value on top of the plot element. If stacked is
also True, annotate the last category. if 'all' and stacked,
annotate all category
full: boolean, default False.
If true, set to full area stacked chart. Only work if stacked is True.
smooth: boolean, default False.
If true, smooth the line.
sort_on: {'index', 'values', int, function}, default 'index'.
If index, sort index on lexicographical order. use as s.sort_index()
if values, sort based on values. Use as s.sort_values()
If function, use it as aggregate (e.g. grouped.agg('mean' or np.mean))
if integer, treat as value and deviate from that value
sort_c_on: string, default None.
specify a category as basis sort value if c is specified. Must be
specified when use sort_on other than default value.
ascending: boolean, default True
sort ascending vs. descending
Returns
-------
Chart Object
"""
return make_chart(df,type='line',x=x,y=y,c=c,how=how,stacked=stacked,
area=area,full=full, smooth=smooth,
sort_on=sort_on, sort_c_on=sort_c_on, ascending=ascending,
annotate='top' if annotate is True else annotate)
def bar_line(df, x, ybar, yline, bar_aggfunc='mean', line_aggfunc='mean',
sort_on='index', ascending=True, is_distinct=False,
hide_split_line=True, style_tooltip=True):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x: string
column to be used as category axis
ybar: string
column to be used as bar values
yline:
column to be used as line values
bar_aggfunc: string (mapping function) or function, default 'mean'
Function to use for aggregating groups on bar values
line_aggfunc: string (mapping function) or function, default 'mean'
Function to use for aggregating groups on line values
sort_on: {'index', 'ybar', 'yline'}, default 'index'
sorting x-axis. If index, sort on x. if either `ybar` or `yline`,
sort based on values
ascending: boolean, default True
sort ascending vs. descending
is_distinct: boolean, default False
Don't use aggregation on this data. Will use drop_duplicates instead.
Ignore `bar_aggfunc`, `line_aggfunc`, `sort_on`, `ascending` parameters.
sort_on deliberately disabled in is_distinct mode to allow already
sorted distinct data.
hide_split_line: boolean, default True
Whether to hide the split line of both y-axis.
style_tooltip: boolean, default True
Whether to offer help to style tooltip. If True, execute
`chart.set_tooltip_style(trigger='axis',axis_pointer='shadow')`
Returns
-------
Chart Object
"""
if sort_on not in ['index', 'ybar', 'yline']:
raise ValueError("Invalid Parameters")
return make_chart(df, x=x, ybar=ybar, yline=yline,
bar_aggfunc=bar_aggfunc, line_aggfunc=line_aggfunc,
is_distinct=is_distinct,
sort_on=sort_on, ascending=ascending,
hide_split_line=hide_split_line,
style_tooltip=style_tooltip,
c=None, type='bar_line')
def hist(df,
x,
c=None,
bins=10,
normed=False,
stacked=False,
annotate=None,
density=False):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x: string
columns to be used as category axis
c: string, default None
another grouping columns inside x-axis
bins: int, default 10
Set number of bins in histogram
normed: boolean, default False
Whether normalize the histogram
stacked: Boolean, default False.
Whether to stacked category on top of the other categories.
annotate: string, {'all',True, None} default None
if True, annotate value on top of the plot element. If stacked is also
True, annotate the last category. if 'all' and stacked, annotate all
category
density: boolean, default False.
Whether to add density to the plot
Returns
-------
Chart Object
"""
return make_chart(df,
type='hist',
x=x,
c=c,
bins=bins,
normed=normed,
stacked=stacked,
density=density,
annotate='top' if annotate is True else annotate)
def line_tidy(df,
full=False,
stacked=False,
area=False,
annotate=False,
smooth=False):
"""
This plot assume DataFrame can be directly consumed (tidy data). Used for
customized manipulation data that normal plot can't provides.
* data is only 2-dimension. There is no hierarchical columns or index.
* data is cleaned and aggregated.
* No duplicate values for each index-column pair (tidy)
* index is used for category x-axis.
* each column corresponds to one series.
* values in one column belong to column data series.
df: pd.DataFrame
data to be used for the chart
stacked: boolean, default False.
Whether to stacked category on top of the other categories.
area: boolean, default False.
Whether to fill the area with line colors.
full: boolean, default False.
If true, set to full area stacked chart. Only work if stacked is True.
annotate: string, {'all', True, None} default None
if True, annotate value on top of the plot element. If stacked is
also True, annotate the last category. if 'all' and stacked,
annotate all category
smooth: boolean, default False.
If true, smooth the line.
Returns
-------
Chart Object
"""
return make_chart(df, x=df.index.name, c="unnamed",
stacked=stacked, area=area, full=full,
type='line_tidy', smooth=smooth,
annotate='top' if annotate is True else annotate)
def scatter(df, x, y, s=None, c=None, saturate=None, size_px=(10, 70)):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x,y: string, columns in pd.DataFrame
Used as coordinate in scatter chart
s: string, columns in pd.DataFrame default None
Used as sizing value of the scatter points
c: string, default None
column used as grouping color category
saturate: string, default None
column to use for saturation
size_px: tuple, default (10,70)
boundary size, lower and upper limit in pixel for min-max scatter points
Returns
-------
Chart Object
"""
return make_chart(df,type='scatter',x=x,y=y,s=s,c=c,
saturate=saturate,size_px=size_px)
def hist(df,
x,
c=None,
bins=10,
normed=False,
stacked=False,
annotate=None,
density=False):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x: string
columns to be used as category axis
c: string, default None
another grouping columns inside x-axis
bins: int, default 10
Set number of bins in histogram
normed: boolean, default False
Whether normalize the histogram
stacked: Boolean, default False.
Whether to stacked category on top of the other categories.
annotate: string, {'all',True, None} default None
if True, annotate value on top of the plot element. If stacked is also
True, annotate the last category. if 'all' and stacked, annotate all
category
density: boolean, default False.
Whether to add density to the plot
Returns
-------
Chart Object
"""
return make_chart(df,type='hist',x=x,c=c,bins=bins,normed=normed,
stacked=stacked, density=density,
annotate='top' if annotate is True else annotate)
def read_df(self, df):
"""
Similar to resync_data, but the difference is creating new Chart object instead of replot
on same cell
Parameters
----------
df: pd.DataFrame
"""
# TODO: make data_columns only just required, except scatter.
if (self._kwargs_chart_['data_columns'] != df.columns).all():
raise AssertionError(
'New data columns is not identical with existing one')
elif df.values.shape[0] == 0:
raise AssertionError('Empty DataFrame')
copy_chart = deepcopy(self)
from krisk.plot.make_chart import make_chart
sub_chart = make_chart(df, **copy_chart._kwargs_chart_)
copy_chart.option = sub_chart.option
copy_chart._chartId = sub_chart._chartId
return copy_chart
def waterfall(s, color_coded=False, annotate=None,
up_name='positive', down_name='negative'):
"""
Firxt category axis automatically not float.
Parameters
----------
s: pd.Series
color_coded: boolean, default to False
Whether to color coded negative and positive values
annotate: {'inside', 'outside', None}, default to None
annotate the bar with actual value.
up_name,down_name: string
increase/decrease bar name. Only work if color_coded is True
Returns
-------
Chart Object
"""
if annotate not in [None, 'inside', 'outside']:
raise ValueError("invalid parameters")
return make_chart(s, type='waterfall', up_name=up_name, down_name=down_name,
color_coded=color_coded, annotate=annotate)
def read_df(self, df):
"""
Similar to resync_data, but the difference is creating new Chart object instead of replot
on same cell
Parameters
----------
df: pd.DataFrame
"""
# TODO: make data_columns only just required, except scatter.
if (self._kwargs_chart_['data_columns'] != df.columns).all():
raise AssertionError(
'New data columns is not identical with existing one')
elif df.values.shape[0] == 0:
raise AssertionError('Empty DataFrame')
copy_chart = deepcopy(self)
from krisk.plot.make_chart import make_chart
sub_chart = make_chart(df, **copy_chart._kwargs_chart_)
copy_chart.option = sub_chart.option
copy_chart._chartId = sub_chart._chartId
return copy_chart
def line(df,
x,
y=None,
c=None,
how=None,
stacked=False,
area=False,
annotate=None,
full=False,
smooth=False,
sort_on='index',
sort_c_on=None,
ascending=True):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x: string
columns to be used as category axis
y: string, default None
if None, use count of category. otherwise aggregate based on y columns
c: string, default None
category column inside x-axis
how: string, default None
to be passed to pd.group_by(x).aggregate(how). Can be mean,median,
or any reduced operations.
stacked: boolean, default False.
Whether to stacked category on top of the other categories.
area: boolean, default False.
Whether to fill the area with line colors.
annotate: string, {'all',True} default None
if True, annotate value on top of the plot element. If stacked is
also True, annotate the last category. if 'all' and stacked,
annotate all category
full: boolean, default False.
If true, set to full area stacked chart. Only work if stacked is True.
smooth: boolean, default False.
If true, smooth the line.
sort_on: {'index', 'values', int, function}, default 'index'.
If index, sort index on lexicographical order. use as s.sort_index()
if values, sort based on values. Use as s.sort_values()
If function, use it as aggregate (e.g. grouped.agg('mean' or np.mean))
if integer, treat as value and deviate from that value
sort_c_on: string, default None.
specify a category as basis sort value if c is specified. Must be
specified when use sort_on other than default value.
ascending: boolean, default True
sort ascending vs. descending
Returns
-------
Chart Object
"""
return make_chart(df,
type='line',
x=x,
y=y,
c=c,
how=how,
stacked=stacked,
area=area,
full=full,
smooth=smooth,
sort_on=sort_on,
sort_c_on=sort_c_on,
ascending=ascending,
annotate='top' if annotate is True else annotate)
def bar_line(df,
x,
ybar,
yline,
bar_aggfunc='mean',
line_aggfunc='mean',
sort_on='index',
ascending=True,
is_distinct=False,
hide_split_line=True,
style_tooltip=True):
"""
Parameters
----------
df: pd.DataFrame
data to be used for the chart
x: string
column to be used as category axis
ybar: string
column to be used as bar values
yline:
column to be used as line values
bar_aggfunc: string (mapping function) or function, default 'mean'
Function to use for aggregating groups on bar values
line_aggfunc: string (mapping function) or function, default 'mean'
Function to use for aggregating groups on line values
sort_on: {'index', 'ybar', 'yline'}, default 'index'
sorting x-axis. If index, sort on x. if either `ybar` or `yline`,
sort based on values
ascending: boolean, default True
sort ascending vs. descending
is_distinct: boolean, default False
Don't use aggregation on this data. Will use drop_duplicates instead.
Ignore `bar_aggfunc`, `line_aggfunc`, `sort_on`, `ascending` parameters.
sort_on deliberately disabled in is_distinct mode to allow already
sorted distinct data.
hide_split_line: boolean, default True
Whether to hide the split line of both y-axis.
style_tooltip: boolean, default True
Whether to offer help to style tooltip. If True, execute
`chart.set_tooltip_style(trigger='axis',axis_pointer='shadow')`
Returns
-------
Chart Object
"""
if sort_on not in ['index', 'ybar', 'yline']:
raise ValueError("Invalid Parameters")
return make_chart(df,
x=x,
ybar=ybar,
yline=yline,
bar_aggfunc=bar_aggfunc,
line_aggfunc=line_aggfunc,
is_distinct=is_distinct,
sort_on=sort_on,
ascending=ascending,
hide_split_line=hide_split_line,
style_tooltip=style_tooltip,
c=None,
type='bar_line')