napjon/krisk

Statistical Interactive Visualization with pandas+Jupyter integration on top of Echarts.

Overview

Krisk bring Echarts to Python Data Science Ecosystem, and helpful tools for high level statistical interactive visualization.

Dependencies

• Python 3.5 (Python 2.7 should be supported, haven't test it yet)
• Jupyter Notebook 4.2.x
• Pandas 0.18.x
• Echarts 3.2.1 (built-in)

Install

pip install jupyter pandas krisk
jupyter nbextension install --py krisk --sys-prefix
jupyter nbextension enable  --py krisk --sys-prefix

Tutorials

• Introduction
• Themes and Colors
• Legend, Title, and Toolbox
• Resync Data and Reproducible Charts
• Declarative Visualization
• Waterfall and Barline Chart
• Tidy Plot: Time Series and Other Custom Data Manipulation

What It Does

• Chart Integration with Jupyter Notebook, widgets, and Dashboard.
• Statistical interactive visualization
• Ability backed by Echarts (Toolbox, Transition, Tooltip, etc.)

What It Doesn't Do

Krisk won't implement all features of Echarts. For more advanced usage, Krisk users can use JSON option (or HTML) output produced by Krisk to optimize in Javascript.

Only basic charts are supported for explanation visualization. The plan will support:

• More complex line, bar, scatter, and histogram.
• Geoscatter plot
• Time Series

Of course, contributions are welcome to support all chart types and advanced features.

Motivation for Another Visualization Library

Krisk is targeted for building interactive dashboard application on top of two key components of Jupyter framework, ipywidgets and Jupyter Dashboard.

Krisk is also act as tool to support reproducible chart by utilizing pandas DataFrame as data input.

How to Contribute

To contribute and unit tests your changes, please do the following,

1. Fork this repository
2. Clone this repo and do unit test,

pip install coverage pytest
git clone https://github.com/your-username/krisk.git
cd krisk
coverage run --source krisk -m py.test

License

New BSD

krisk's Issues

Add GeoScatter feature

Move gallery notebooks to nbviewer

Add Support for Python 2.7

Add legend, toolbox, and tooltip to rcParams

Add Coverage Testing

Create notebooks tutorial

Refactor plot.py to its own package

Add Bar Mode for Sequential, Diverging, and Qualitative

add 'mode' as keyword parameter for bar chart.

• If None, treat the bar mode as "Qualitative", and sort on lexicographical order.
• If Boolean, treat the bar as "Sequential", whether the order sort ascending.
• If string, either median or mean and deviation from that value
• if integer, treat as value and deviate from that value

Add bar_line chart

Adding chart with bar and line with different series but same x-axis. For example,

kk.bar_line(df, x, y_bar, y_line, c=None, .....)

Line Plot should have default trigger axis and axis pointer shadow

Add trendline parameters for bar chart

Add trendline that shows changes for bar chart

Add Waterfall chart

Support Waterfall.
Echarts Examples:

• http://echarts.baidu.com/demo.html#bar-waterfall
• http://echarts.baidu.com/demo.html#bar-waterfall2

Have the actual data place outside of the option

Don't embed the actual data in the option variable. Instead, assign variable for each of category, and let data option reference to that variable

Update tests

Add Time Series Support

Added Declarative Visualization to Krisk

Krisk users can just specify JSON option of Echarts to plot directly in Jupyter

Not support nbconvert

Reproduce:

1. Use Intro.ipynb
2. Remove from krisk import init_notebook; init_notebook() since it's not gonna publish to nbviewer.
3. Use nbconvert to export current page as html

Output:
Javascript part failed to print out.
Console error:
require.min.js:8 Uncaught Error: Script error for: dark
require.min.js:8 Failed to load resource: net::ERR_FILE_NOT_FOUND
...

Expected Output:
Charts should be shown as it run in notebook after exporting.

Possible solution:
It seems like after exporting, it still tries to retrieve source js files from local server and current folder. Is that possible we can define the source files' location as commands in python?

ENV:
Python3.5
Jupyter Notebook 4.2.1
Echarts 3.2.1 (built-in)

if category, bar plot not shown all index values

See http://nbviewer.jupyter.org/github/napjon/krisk/blob/master/notebooks/themes-colors.ipynb

Not all year index values shown

Set Style for y and x axis

Support for Pie Chart

As far as I know, the current package only support line,scatter,bar,bar_line types...Thanks...

Resync Data also have to include external option

There is option that generated by make_chart, but there are also additional option provided by setter method. Resync Data have to handle that as well.

Add smooth line for histogram and all type of line chart as well

Echarts Density Line for Histogram: http://echarts.baidu.com/demo.html#mix-line-bar

• Add an option to disable bar, line, or using both for histogram.
• For all type of line chart, added smooth boolean as additional argument, add to every line series.

Add versioneer

Add Saturation for Scatter Plot

Error in Pip Install

I'm having a similar problem to #70 **
Collecting krisk
Using cached krisk-0.3.0.tar.gz
Complete output from command python setup.py egg_info:
Traceback (most recent call last):
File "", line 1, in
File "/private/var/folders/6s/fdz619ns34jfkhhjfqmwb3br0000gn/T/pip-build-x99MZJ/krisk/setup.py", line 2, in
from krisk import version
File "krisk/init.py", line 5, in
import krisk.plot.api as plot
File "krisk/plot/api.py", line 1, in
from krisk.plot.make_chart import make_chart
File "krisk/plot/make_chart.py", line 2, in
from krisk.chart.api import Chart
File "krisk/chart/api.py", line 1, in
from krisk.chart.core import rcParams, Chart
File "krisk/chart/core.py", line 189
SyntaxError: Non-ASCII character '\xef' in file krisk/chart/core.py on line 189, but no encoding declared; see http://python.org/dev/peps/pep-0263/ for details**

The last update doesn't seem to fix this issue.

Add CircleCI Integration

FileNotFoundError: [Errno 2] No such file or directory: 'C:\\Program Files\\Anaconda3\\lib\\site-packages\\krisk\\static/krisk.js'

hi there

python3.5
win10
pip install krisk

i run this page code， http://nbviewer.jupyter.org/github/napjon/krisk/blob/master/notebooks/Intro.ipynb

when i run , kk.bar(df,'continent')

i got error

---------------------------------------------------------------------------
FileNotFoundError                         Traceback (most recent call last)
C:\Program Files\Anaconda3\lib\site-packages\IPython\core\formatters.py in __call__(self, obj)
    309             method = get_real_method(obj, self.print_method)
    310             if method is not None:
--> 311                 return method()
    312             return None
    313         else:

C:\Program Files\Anaconda3\lib\site-packages\krisk\chart\core.py in _repr_javascript_(self)
    574                                       width=self._size['width'],
    575                                       height=self._size['height']))+\
--> 576                 (self._get_resync_option_strings(self.option))
    577 
    578     def _get_duplicated(self):

C:\Program Files\Anaconda3\lib\site-packages\krisk\chart\core.py in _get_resync_option_strings(self, option)
    553         """Resync Chart option"""
    554 
--> 555         js_template = get_content(JS_TEMPLATE_PATH)
    556         event_template = get_content(EVENT_TEMPLATE_PATH)
    557 

C:\Program Files\Anaconda3\lib\site-packages\krisk\util.py in get_content(filepath)
     13         """Retrieve content from file"""
     14         abs_path = join_current_dir(filepath)
---> 15         with open(abs_path, 'r') as f:
     16                 return f.read()
     17 

FileNotFoundError: [Errno 2] No such file or directory: 'C:\\Program Files\\Anaconda3\\lib\\site-packages\\krisk\\static/krisk.js'

Out[4]:
<krisk.chart.core.Chart at 0x27540c21ba8>

Add recursive set option

Consistent API parameters with Pandas and Matplotlib (BREAKING API)

This will break API change.

• Convert size to s
• Convert category to c

Use nbextension for offline JS

Add full area (100%) for stacked bar and line chart

Use None instead of 0 for missing value

0 treated as value, and thus makes line chart dropped to zero because of missing data.

Add formatting option

Now numbers rendered as is, and it doesn't do formatting properly when go as high as hundreds of millions. Should add additional parameters to control how numbers are rendered.

Error in pip install

Collecting risk
Using cached krisk-0.3.0.tar.gz
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):
      File "<string>", line 1, in <module>
      File "/private/var/folders/c4/yxk5p2k93z53gw5ywq750x3nhlvp8p/T/pip-build-imEuXs/krisk/setup.py", line 2, in <module>
        from krisk import __version__
      File "krisk/__init__.py", line 5, in <module>
        import krisk.plot.api as plot
      File "krisk/plot/api.py", line 1, in <module>
        from krisk.plot.make_chart import make_chart
      File "krisk/plot/make_chart.py", line 2, in <module>
        from krisk.chart.api import Chart
      File "krisk/chart/api.py", line 1, in <module>
        from krisk.chart.core import rcParams, Chart
      File "krisk/chart/core.py", line 189
    SyntaxError: Non-ASCII character '\xef' in file krisk/chart/core.py on line 189, but no encoding declared; see http://python.org/dev/peps/pep-0263/ for details```

Add TimeDelta Support

In Clinical Trial, there are various cases where start date for each patient is different. We have to make baseline date for each patient and measure compliance rate. Timedelta operations can benefit case like this.

Set global config for all charts in current session

Users don't have to set config for each chart. Currently targeting for:

• Themes, Color, and Pallette
• Size

Break unit test into smaller pieces instead of comparing chart option to JSON output

how export .ipynb to html,I use command to export the chart can not show

i use command like this :
ipython nbconvert --to html --template fill mynotebook.ipynb
the question is all the krisk img can not show ....