In [1]:
%run talktools
Plotly
collaborative, interactive
and online plotting with $\TeX$
What is Plotly?¶
Plotly is an online analytics and data visualization tool.
In [2]:
from IPython.display import IFrame
IFrame("https://plot.ly/~Dreamshot/407", width="900px", height='500px')
Out[2]:
Why is Plotly at TUG 2014?¶
- You can use $\TeX$ symbols to annotate Plotly graphs, (using the MathJax display engine)
In [3]:
from IPython.display import IFrame
IFrame("https://plot.ly/~PlotBot/3", width="900px", height='500px')
Out[3]:
But really ... Why is Plotly at TUG 2014?¶
... Philosophy¶
$\TeX$ provides a system that gives exactly the same results on all computers
- $\TeX$ is cross-operating system
$\TeX$ allows anybody to produce high-quality documents efficiently
- $\TeX$ is a free and cross-text editor platform
Plotly applies the same core principles to graphs
- Plotly is free, cross-operating system and cross-scientific computing language
Collaboration in data-intensive fields sometimes feels like this:¶
Plotly solves this collaboration (i.e. reproducibility) problem:
- Plotly graphs are closely connected to their underlying data (more later)
- Plotly graphs are stored in the cloud
- Plotly provides a common graphing platform for Python, MATLAB, R, Node.js, Julia and Excel users.
A Plotly graph made in Python¶
In [4]:
from IPython.display import IFrame
IFrame("https://plot.ly/python/histograms/#Overlaid-Histgram", width="1100px", height='520px')
Out[4]:
The same Plotly graph, now made in MATLAB¶
In [5]:
from IPython.display import IFrame
IFrame("https://plot.ly/matlab/histograms/#Overlaid-Histgram", width="1100px", height='520px')
Out[5]:
... and again the same Plotly graph, now made in R¶
In [6]:
from IPython.display import IFrame
IFrame("https://plot.ly/r/histograms/#Overlaid-Histgram", width="1100px", height='520px')
Out[6]:
What if I already have plot-generating code written?¶
That's OK.
Our libraries come with figure converters allowing
- MATLAB,
- matplotlib and
- ggplot2
figures to be converted to Plotly figures with one line of code!
In [7]:
from IPython.display import IFrame
IFrame("https://plot.ly/matplotlib/", width="1100px", height='520px')
Out[7]:
How to make Plotly graph using our web app¶
- Have some data to plot
- Go to plot.ly
Plotly rhymes with collaboration and reproducibility¶
Plotly allows you to retrieve a figure's underlying JSON object!
For example, in Python:
In [8]:
import plotly.plotly as py
fig = py.get_figure("https://plot.ly/~etpinard/450")
fig
Out[8]:
Remake this plot, with a few modifications¶
In [9]:
import plotly.plotly as py
fig = py.get_figure("https://plot.ly/~etpinard/450") # as in last slide
# Modify the title
fig['layout'].update(title="The Historical Population of Portland, OR")
# Modify the y-axis label
fig['layout']['yaxis'].update(title="Population")
# Plots the data with marker points (not line)
fig['data'][0].update(mode='markers')
# Re-generate plot, get a unique URL
py.plot(fig, filename="tug-conf-example")
Out[9]:
Go to plot's URL (or use py.iplot() to embed plot in IPython notebook)
Save a copy on your computer¶
In [10]:
import plotly.plotly as py
py.image.save_as(fig, filename='test.png')
In [11]:
from IPython.display import Image
Image("test.png")
Out[11]:
Moreover, Plotly¶
- is also a social network (with a twitter-like feed of figures and commenting on each graphs)
- allows users to make private figures (kind of like github, see our plans)
- allows users to make streaming plots (e.g. a never-ending double pendulum simulation: graph and code)
- is developing of Open Source Libraries (e.g Python, MATLAB, R)