Interactive Dashboard

The python visualisation landscape is rich and diverse. But also, given the diversity of the tools - it is hard for vis libraries, it has a lot of challenges in learning libraries and varying degree of maturity for the tools.

See this PyCon 2017 Talk from JakeVandeplas on the Landscape

Speakerdeck link: https://speakerdeck.com/jakevdp/pythons-visualization-landscape-pycon-2017
Youtube link: https://www.youtube.com/watch?v=FytuB8nFHPQ
Github link: https://github.com/jakevdp/python-vis-landscape

We will use a tool called facet-dive for multidimensional exploration of data

Installation Instruction:

git clone https://github.com/PAIR-code/facets
cd facets
conda install protobuf
jupyter notebook --NotebookApp.iopub_data_rate_limit=10000000



In [7]:

    
import pandas as pd



In [10]:

    
df = pd.read_csv("../data/default.csv")



In [11]:

    
df.head()



In [18]:

    
jsonstr = df.to_json(orient='records')



In [20]:

    
# Display the Dive visualization for this data
from IPython.core.display import display, HTML

HTML_TEMPLATE = """<link rel="import" href="/nbextensions/facets-dist/facets-jupyter.html">
        <facets-dive id="elem" height="600"></facets-dive>
        <script>
          var data = {jsonstr};
          document.querySelector("#elem").data = data;
        </script>"""
html = HTML_TEMPLATE.format(jsonstr=jsonstr)
display(HTML(html))



In [ ]:

	default	amount	grade	years	ownership	income	age	y_pred	y_proba
0	0	1000	B	2.0	RENT	19200.0	24	0	0.15
1	1	6500	A	2.0	MORTGAGE	66000.0	28	1	0.80
2	0	2400	A	2.0	RENT	60000.0	36	0	0.08
3	0	10000	C	3.0	RENT	62000.0	24	0	0.05
4	1	4000	C	2.0	RENT	20000.0	28	1	0.82

Interactive Dashboard

Facet-Dive: Interactive Exploration of Data and Classification