Presentation Objectives:
Introduce you to a the new approach for BI
Get hands on with Visual Analytics
Understand how Visual Statistics extends the capabilities
Learn how to get started with Visual Analytics
Answer your questions
Hello everyone and Welcome to SAS’s Visual Analytics Hands-On Workshop, we are very happy to have you here. My name is ______________ and I am (position and title). I would also like to introduce you to (other SAS people in the room, i.e. Shawn and Ron/ Mark ) (Names, Titles, etc.). We are all thrilled that you will be joining us this morning (afternoon) for this interactive workshop.
This is already started with Python programmers, etc. They did not wait for it… they went and created this streamlined process on their own. So how to address that crowd. The new data scientists have been doing this parallel process, with their own tools already. The problem is that it is often not organized, enterprise-wide, governed. SAS can help move it from ‘experimental status’ to productionized environment.
Governed visualization – not desktop, able to secure and you have control
Data Sandbox – Eliominate the proliferation of data to desktops this is a place where new/not yet corporate data can comingle with enterprise data
Unguided analysis – Can look at any data and find your answers in a free form way
Data autonomy – users can create variables, hierarchies and complex calculations
Leverage SAS – ability to expose SAS information to users in a more consumable way, less time spent on mundane data presentation
Analytic insight – ability to do simple statistical methods without the need to know stats. Forecasting, decision trees movement to VS
Data Prep – We allow users to do the full data lifecycle in a secure, managed and shared way. DI metadata sharing etc.
Scalability – Ability to scale and deliver value to the entrprise
SAS Visual Analytics is comprised of these key components…
Visual Data Builder
Visual Analytics Administrator
Visual Analytics Explorer
Visual Analytics Designer
Mobile BI
The SAS LASR Analytic Server
And a streamlined, action-oriented central entry point for the key capabilities is provided by the “Hub”.
Let’s take a closer look at each of these components…
SAS Visual Analytics Explorer - Organizations have realized the importance of analysing every possible aspect of their data, the need of analytically exploring any size of data and understand the patterns, trends more effectively going beyond tabular reports.
SAS Visual Analytics brings a highly visual data exploration interface allowing users to take advantage of SAS predictive analytics power to gain insights from their data, with simple to use user actions, and surface consumable analytical results in a visual format, helping customers find relationships and discrepancies in their data. As part of the integrated infrastructure, users can share their findings to web and mobile users.
One must note that visual exploration is about using advanced analytics to *visually explore* any size of data and it is not a reporting tool.
Here is a sample screenshot of the exploration environment. The interface provides options for selected graphics as well as auto-charting capabilities, drag-n-drop environment for generating visualizations, interactive and dynamic filtering, ability to create dynamic hierarchies by the end-users without the need for pre-defined dimensional structures.
SAS Visual Analytics Designer – a component that brings the capabilities of classic reporting and highly visual dashboarding as part of single report. The designer creates reports out of various visuals such as graphs, tables, gauges, prompts, geo maps, texts, and images with the ability to use multiple data sources as part of the single report.
Users can set various types of interactions across the report objects in a WYSIWYG design format, derive new data items, create hierarchies, add comments, export data. Reports created in this interface (or exported from the Explorer interface) are readily available for users to access via browser or supported mobile devices.
Here is a sample screenshot of the designer with various reporting elements, in a precision layout. Each visual element could have come from different data sources.
SAS Mobile BI - SAS Visual Analytics provides SAS’ native mobile application allowing users to view their reports and dashboards on their selected devices. Currently supported devices are iPad and Android-based tablets. SAS is continuously monitoring the market trend and the demand from customers for other operating systems like Windows and Blackberry and will be extending support for new devices in future releases as required.
SAS Mobile BI provides adaptive presentation so that users will not have to create separate reports for each device type. With this offering, SAS Mobile BI provides all the required capabilities via a highly visual and interactive mobile application backed by centralised metadata security. While viewing the SAS Reports, the content will be downloaded to the device and thus fully support an interactive offline analysis. Administrators can force tethering, where mobile app will work only when it is connected to the server. When the report is closed, the data is wiped out.
SAS Visual Statistics 7.1
Product Release Date/Month: Oct. 2014
Contact: Tapan Patel
Last Updated: October, 2014
Irrespective of big data or large data, every analytics project should go through the iterative analytics (data to decision) lifecycle. Typically four steps involved are: manage/prepare data, explore/visualize, model and deploy & monitor.
The role of SAS Visual Statistics is to (primarily) address the data exploration and interactive model development stages of the analytics lifecycle.
It allows customers to understand on why certain events, outcomes happen and what are the key relationships. Users ask for more interactions from the data, demand drill-down, etc. to identify the root cause and use the information to build predictive models.
It allows customers to build and refine predictive models to assess a future outcome and explain what will happen? For example, is the transaction fraudulent or not or to assess future risk of repayment or how risky is the portfolio given certain conditions in future? Users can dynamically see the impact of changing model properties/parameters and fine tune the model to arrive at the desired results.
Superior interactivity and Rapid Model Building and Refinement with software. What does it mean and how does that help?
In-memory analytic computations dramatically shortens the time required to build and adjust models until you are confident, including the “prototype-review-refine” iterative cycle. It supports the “what you think is what you get” paradigm.
In-memory processing…..Users can instantly see the impact of changing model settings – like adding new variable or removing outliers or changing parameters….
Interactive analytics…Designed for multiple passes through data for multiple analytical jobs. Or a change in a predictive model, does not requires them to repeat the entire process.
Build numerous models by segments/groups instantly from a decision tree or clustering analysis (i.e. stratified modeling) to find new opportunities and take well-qualified decisions..
Predictive analytics can be applied to variety of business issues across all industries and solutions areas….some of the sample use cases listed here could be assigned into common areas like:
Clustering is the task of segmenting a heterogeneous population into a number of more homogenous subgroups. Segmentation does not rely on predefined classes or examples. The records are grouped together on the basis of self-similarity. Clustering is often done as a prelude to some other form of data mining. For example, market segmentation – cluster of customers with similar buying habits and find out which promotion would work best.
Classification deals with prediction of discrete outcomes and arrive at Yes/No decisions with confidence. In business world, many decisions are binary in nature - Churn/No churn, Fraud/No Fraud, Credit Extension or Not, Will you respond to marketing campaign or not. Multi-level classification is similar except that that are more than two level - credit applicants as low, medium or high risk. Mobile vendor is interested in voluntary churn, involuntary churn, an active customer.
Estimation or interval prediction is another form of prediction wherein target level is continuous (i.e. individual records are rank ordered). Examples – insurance firm predicting claim frequency, claim severity and pure premium based on your specific information, predicting box office receipts by segments, etc.
Groups can be specified in source data or be the result of an analytical model (i.e. a cluster) for generating multiple models by segments/groups. The goal is to exploit relationships to the group.
Setup regression analysis using both categorical and continuous business drivers
Work with classification modeling using logistic regression and decision trees
Data driven segmentation using clustering. SAS Visual Statistics lets you visually explore and evaluate segments for further analysis using k-means clustering, scatter plots and detailed summary statistics.
Concurrently build numerous models and process results for each group or segment without having to sort or index data each time. The grouping variables, or their properties, can change from one action to the next and groups are processed without reordering the data. This means more results can be generated for each group on the fly without additional processing overhead.
Governed visualization – not desktop, able to secure and you have control
Data Sandbox – Eliominate the proliferation of data to desktops this is a place where new/not yet corporate data can comingle with enterprise data
Unguided analysis – Can look at any data and find your answers in a free form way
Data autonomy – users can create variables, hierarchies and complex calculations
Leverage SAS – ability to expose SAS information to users in a more consumable way, less time spent on mundane data presentation
Analytic insight – ability to do simple statistical methods without the need to know stats. Forecasting, decision trees movement to VS
Data Prep – We allow users to do the full data lifecycle in a secure, managed and shared way. DI metadata sharing etc.
Scalability – Ability to scale and deliver value to the entrprise