An Introduction to R Jupyter (Azure) Notebooks
by Bryan Cafferky
SQL Saturday in Smithfield, RI
December 9th
http://www.sqlsaturday.com/694/EventHome.aspx
Use Discount
Pre Con: R from A to Z Code
https://r_programming.eventbrite.com Microsoft
For 20% Off
About Bryan Cafferky…
• Microsoft Technical Solutions Enabler
• Decades of IT Experience
• Microsoft Data Platform MVP
• Author of Pro PowerShell for Database Developers
• Experienced in health care, insurance, banking, and ecommerce
• Founded and lead PASS Chapter The RI Microsoft BI User Group and The SNE R User Group
• Proud to be a Nerd.
[email protected]
www.sql-fy.com
https://www.linkedin.com/in/bryancafferky
@BryanCafferky – follow me!
Time to
Find Your
Inner
Child!
What is Jupyter Notebook?
When to Use Notebooks?
Teaching Collaboration
Research Exploration
Why Use Notebooks?
75+ Languages Supported Sharing
Jupyter Notebook
Open Source
Written in Python – originally called IPython
Real time execution and rendering.
Many languages supported.
Supports Spark API (pySpark, SparkR)
Can share via GitHub, JupyterHub, and Azure.
Built-in Viewer Support in GitHub.
Common Funding = Common Benefit
It’s not
just for
Python
Anymore!
Who uses Jupyter Notebooks?
What’s in a name?
Emphasizes
Jupyter
multi-language
support.
R
Julia
Python
Why Notebook?
From Galileo’s
Notebook –
Drawing’s of
Jupiter’s
satellites.
Kernel Architecture
JSON format
Can be many kernels:
• Python
• R
• Julia
• Ruby
• Perl
• PowerShell
http://jupyter.readthedocs.io/en/latest/architecture/how_jupyter_ipython_work.html
Flexible Notebook Conversions
nbconvert
Formats include:
• HTML Slides
• PDF
• Markdown
• Script
http://jupyter.readthedocs.io/en/latest/architecture/how_jupyter_ipython_work.html
What is Jupyter Notebook?
Execute the cell
Command Prompt
called a cell
• Code executes in real
time.
Rich visualizations
rendered inline. • Can go back, edit and re-
run cells.
• Like a cross between a
command line and Excel
spreadsheet.
Starting Jupyter Notebook
Start the command shell.
Installing Jupyter Notebook
Start the Jupyter Notebook
Windows Command Prompt
Note: The notebook starts pointing to the folder from where the notebook was started.
The Jupyter Server Running
A First Look at a Notebook
Select New and the
kernel, i.e. R
A First Look at a Notebook
Kernel Running
Enter R code and
press Shift + Enter
Or
Click on
The Cell Toolbar
Save Move Cell Stop Cell
Cut Cell Copy Cell
Notebook Down 1 Execution
Run Code Set Cell
Paste Cells Move Restart the
Insert Cell in Selected Content
Below Cell(s) Up 1 Kernel
Cell(s) Type
• New notebook
• Sample notebooks
• Python
• R
• Cell Magic
• Widgets
Exploring
Jupyter…
http://jupyter.org/
• Documentation
• Free notebook play area
Use R and Python Together… Loads R
Python
Extension
%R to run R Default
code in the language is
cell Python 3
R ggplot call
passing data
frame, df, from
Python
%%R defines data to
be passed between R
and Python:
-i = Python to R
-o = R to Python
Accessing Spark with pyspark…
Import
pyspark
module
Set the
Resilient Spark
Distributed context
Dataset
Point
context to
HDFS file
Reads the
data in
Spark
Sample the
data
Presentations from Jupyter…
Switch to Organize
Slideshow Slides
Mode
Github Renders Notebooks Automatically…
Advanced Data Science Tool
Architecture
Features
Maximize Notebook Value
https://notebooks.azure.com/
To Install R
Packages
install.packages("psych", repo="http://cran.us.r-project.org")
Jupyter Notebooks on Azure
Expand/
Collapse
Control
Jupyter Notebooks on Azure
• Built-In Support to
maintain
notebooks.
• Share with others.
• Initially supports R,
Python 2 and
Python 3.
Jupyter Notebooks on Azure
Use the Gallery to Get a
Jump Start
Why Azure Notebooks?
Provides Custom Development IDE within the Azure Portal
Notebooks Can Access All Your Azure Resources
Share and Collaborate
Secure
One Tool for Any Language
RStudio Fights Back…
Expand/
Collapse
Control
Output
rendered in
the editor
Installing Additional R Packages
Install Pop Up in RStudio
• Installing additional R Packages
• Find out where R is installing packages.
• Find where Anaconda is installed. R packages are in subfolder \R\library
• Example: C:\Users\BryanCafferky\Anaconda3\R\library
• Copy the package folder from the R folder to the Anaconda folder.
How to get Jupyter Notebook…
Installing Jupyter Notebook
• Install Python – Jupyter is a Python application.
• Recommend the Anaconda distribution which has many useful packages.
• Choose 32 or 64 Bit as appropriate.
• Either Python 2.x or Python 3.x will work. Python 3 is the new standard.
https://www.continuum.io/downloads
• Install R if you do not have it installed. An R IDE is not required.
https://cran.r-project.org/bin/windows/base/
Installing Jupyter Notebook
Install the R Kernel with Extensions, i.e. popular R packages.
Windows Command Prompt
Jupyter Notebook – Wrapping Up
Key Features
Open Source
Written in Python – originally called IPython
Real time execution and rendering
Many languages supported
Supports Spark API
Sharing many ways, i.e. file system, GitHub, JupyterHub, Azure
Built-in Viewer Support in GitHub
Jupyter Notebooks – Wrapping Up
Questions?
