The latest releases of Cognos Analytics starting with, version 11.1.2, comes with a much-awaited inclusion: Jupyter Notebooks. With support for most of the standard data science packages based on Python, Jupyter Notebooks promises to expand the boundaries of data analysis by offering data scientists an environment where they can collaborate and interact with data in real-time.
The Notebooks server arrives readily compatible with Python packages from Anaconda and PixieDust. In essence, Jupyter Notebooks is an amalgamation of two chief components, which operate synergistically to allow data scientists to interact with data, and crucially, share this data over a shared network with fellow data analysts for collaboration purposes.
These components are the Notebook documents and a browser-oriented web application.
What Are Jupyter Notebook Documents?
Notebook documents are the digital representation of the data within the application. It's where raw data turns into images, mathematics, media representations, explanatory texts, and input/output computations. Through Notebooks, you can convert data into equations, visualizations, live code, and narrative texts, and share it within the web application.
Notebooks are equipped with a wide range of features to facilitate data expression. Access to features like data visualization, machine learning, statistical modeling, data transformation, and numerical simulation gives you, the data scientist, a comprehensive suite from which you can fully capture the nature of the data, and share it in its truest form.
The Jupyter Notebooks Web Application
Interacting over the Jupyter Notebooks server would not be possible without the second component, which is a browser-oriented tool that enables the interactive computing of data sets in the formats mentioned above. Here, data scientists can collaborate on documents featuring mathematical expressions, equations, narrative texts, and rich media outputs.
How Jupyter Notebooks Improves Data Analysis
Data science is based on three principles. These principles are necessary for the extraction of conclusions from large and varied sets of data.
- The first principle is exploration, which involves the identification of patterns and sequences from random sets of data.
- The second principle is prediction, where we try to understand unknown values by making informed guesses based on information that we already know.
- Lastly, there is inference, which serves the purpose of questioning and analyzing predictions to verify their accuracy.
Notebooks create an environment where data scientists can interact with data in real-time. That means that you can execute codes within this environment, monitor data processes, and apply modifications—all in real-time!
This live interaction with data is only made better by Jupyter Notebook’s sharing capabilities, where you can create and share the computational output, visualizations, explanatory text, and software codes with other data scientists, making collaboration effortless, and dramatically improving productivity.
There are other benefits of utilizing Jupyter Notebooks and Cognos Analytics as a whole in data science. First, it allows you to work without Silos, which is something that many data scientists using Python are still not able to do. That makes data sharing to the general user population possible.
For the longest time, there was no simple way to share results with users, especially when leveraging ungoverned data, as is the case when using Silos. Cognos manages governed data and allows you to create a library of information that is accessible by all users within the network.
The Improved Functionality of Cognos Analytics 11.1.2 and beyond
Other than eliminating the need for Silos and organizing data sets into libraries that can be accessed by the general user population, Jupyter Notebooks will allow you to perform the following actions once it is fully integrated.
- You will be able to control who can access documents, both in your personal folder (My Content) or in collaborative projects (Team Content). It is possible to assign roles in a manner that offers different clearance levels to various members of your team.
- Notebooks can be saved in either My Content or Team Content.
- You can schedule automatic executions to occur without your supervision.
- Notebooks can be run in the background while other tasks run in the foreground. You can turn on alerts to receive a notification upon completion.
- Jupyter Notebooks supports externally created Notebooks, which you can upload directly to either My Content or Team Content. It will be compatible with documents created in earlier versions of Cognos Analytics.
Jupyter Notebooks Brings the Following Benefits to Data Science:
Jupyter Notebooks promises to improve collaboration amongst data scientists with enhanced functionality. Some of its notable benefits, the one you will most certainly appreciate, are:
- The ability to share notebooks and insights with fellow data scientists
- The ability to leverage governed data from data sets, modules, packages, and uploaded data files
- The ability to avail discoveries to both power users and casual users within the network
- The ability to upload Notebooks created outside Cognos Analytics and save them in the Cognos Content Store
- The ability to include output sources to data in My Content and Team Content
The collaboration of Jupyter Notebooks and Cognos Analytics creates a formidable duo that will undoubtedly improve the experience of data scientists on all tiers of expertise. Learn more about the integration of Jupyter Notebooks into Cognos Analytics Software on our Webinar: What's New in Cognos Analytics: Jupyter Notebooks.