Azure Notebooks at Microsoft Connect() 2018
The Azure Notebooks team is excited to share with you our latest update that we announced at Connect() 2018 during Scott Hanselman’s keynote. Here’s what’s new.
Azure Notebooks is a hosted service for Jupyter Notebooks that requires no installation. It enables you to get started quickly on prototyping, data science, academic research, or learning to program Python or R. You can combine code, markdown, images, videos, graphs, all in one format and easily share with partners.
Best of all, this service is free to use, making it a great environment for:
- Teachers who need a hassle-free Python/R environment for classes,
- Presenters who don't want attendees to spend 45 mins installing software,
- Developers and hobbyist who need a quick coding scratchpad,
- Data scientists who need a full R, Python (Anaconda) environment and don't want to spend the time installing everything.
Documentation for Azure Notebooks is found on https://docs.microsoft.com/azure/notebooks.
In addition to the notebooks in this repository, which you can also find on the Azure Notebooks samples library, there are some good external resources as well:
- Intro to Jupyter notebooks
- Intro to Data Science by Jake Vanderplas
- See the Sample notebooks on the Azure Notebooks front page
- A curated list of interesting Jupyter notebooks
Note: You don't need an account to view notebooks, but you will need a Microsoft account (Outlook, Hotmail, Xbox, etc) to run them. You will be prompted to create an account when needed.
There are many interesting sources of data to use in notebooks. A couple of good resources are:
Note that data must be accessible to the service — you can use Github, Dropbox, OneDrive, Azure blobs, SQLAzure, and more.
- Jupyter.org docs
- Jupyter on YouTube
- File an issue for bugs, feature requests, etc. If you need immediate assistance, ping us at firstname.lastname@example.org.
- Have a cool notebook you want to share? Upload & tweet with
- Are you teacher giving a large class or want to give webinar? Test drive the site and if possible give us a heads up on the date & number of users at email@example.com.
Knowing the context of your issue will help us solve it faster. Please include as much of the following as you can when creating a issue:
- Whether the issue is related to Jupyter Notebooks or the Project feature.
- The Jupyter Kernel (
R, etc) in case of problems with a Jupyter Notebook.
- Time of occurrence.
- Web browser the issue occurred in.
- If the issue occurred once or can be reproduced (with steps to repro it).
- Library ID and Notebook Name.
- Logs (
!cat .nb.logor: New > Terminal,
$ cat .nb.log).
- Usage scenario: teaching a class, doing homework, evaluating results for research.
- Add template to root.
What is Jupyter?
Jupyter (formerly IPython), is a multi-lingual REPL on steroids. Azure notebooks is a free service that provides Jupyter notebooks along with supporting libraries as a service. It means you can just login and use it, no installation or setup is necessary. Instructors can also use the service by easily sharing a notebook. Students get their own private copies that they can run. The service is provided by the Python team @ Microsoft, which is part of the Data Group.
What environments are supported and how can I install packages?
Python 3.6.x and Python 2.7.x (Anaconda3 and Anaconda2 5.3.0), with many additional packages.
- You can install packages in a Python notebook:
!pip install <pkg-name>
- You can also install packages using conda:
!conda install -y --name root <pkg-name>
We also support R (3.4.1 running MRO), and F# (4.1).
- To install packages in R:
See also our full FAQ.