SAS

SAS (Statistical Analysis System) is a software system intended for statistical analysis, through various developments and integrations is now an important tool to support the Data Scientist for business planning, forecasting analysis and decision support. A Business Analytics tool that allows you to solve complex problems regarding operations research, risk management, big data management, visual analysis, clustering, forecasting, machine learning, artificial intelligence, etc.

The basic functions include data access, management and organization to carry out descriptive, predictive, prescriptive and even automated statistical analysis. SAS has also developed Advanced Analytics e AI AI applications that transform raw data into operational information to support decisions made within each organization.

Blue BI we have developed in-depth technical and functional skills on the main SAS solutions and modules:

Blue BI solutions based on SAS systems allow functions such as data warehousing and data mining: we connect the data of the corporate information assets with external data series, in an agile and easy-to-use way, supporting decision-making processes aware, immediate and effective. We make sure that every decision made, at every level of management, has a positive impact in terms of optimization of the resources, cost containment, increase in turnover and development of innovative strategies based on hypothesized scenarios.

SAS technology

SAS Viya is the latest SAS architecture, cloud native and continuously updated. This means that once installed, the new features will be automatically downloaded and added to the customer installation. It is also scalable by design, both for data and for the number of users. Thanks to this, it will always adapt as best as possible to the needs. The heart of this architecture is SAS Cloud Analytic Services (CAS). It is an in - memory engine that provides the run - time environment for data management and analytics with SAS.

Architecture

In terms of capacity, the SAS®Viya Platform enables many features, such as:

Data access; data preparation, data quality & information cataloging
Advanced flow steps & information governance
Visualization & reporting
Conversational AI & chatbots
Statistics
Machine learning & deep learning
Model deployment & management
Forecasting
Text analytics
Optimization

The SAS platform enables companies to address business demands by removing the barriers created by large data sizes, data variety and diversity, scalability, integration and greater analytical depth.

Also using and integrating programming languages such as Python, Java and R and open source or third party platforms and REST services. The SAS Viya platform even supports public RESTful APIs hosted in the Cloud.

SAS Viya Key Features

Statistic

Users can explore data and interactively build predictive analytical models. Datascientists, statisticians, and analysts can cooperate and iteratively refine models for each segment or group to make understanding-based decisions.

The models can be developed in the SAS language of SAS Studio or in other languages such as R, Python, Java etc …

Some visuals that can be used:

Artificial Intelligence

The in-memory processing environment can be extended to modern statistical, data mining and machine learning techniques.

In addition to the automatic node modeling functions, there are a number of optimizations and auto-tuning procedures for different machine learning algorithms (Decision Trees, Random Forests, Gradient Boosting, etc.) that improve the performance of the model.

You can also combine structured and unstructured data into machine learning algorithms. By running SAS code or by embedding open source code or by accessing templates from a shared repository.

Forecasting

quickly and automatically produce a large number of reliable forecasts from historical data.

Create projects using visual flow charts with a variety of date, test, historical, real time sets. With the ability to choose the sample model based on the results.

Large-scale time series can be modeled and predicted effectively thanks to a highly parallel and distributed architecture. This essentially satisfies the speed and scalability needed to build models and generate predictions for millions of time series. Massive parallel processing is one of the key benefits for predicting large time series.

You can perform the following activities:

Performing automatic model and variable selection from source data.
Generating predictions automatically using integrated modeling strategies.
Creating your own modeling strategies.
Making hierarchical predictions.
Visually analyzing and diagnosing time series data.
Creating attribute-based filters to view subsets of historical and forecast data.
Replacing predictions using attribute-based filters.
Disaggregate overrides using an optimization model with automatic conflict resolution or manual user resolution.
Exporting projects as SAS or Python code for processing in a batch environment.
Supporting for open source languages (includes External Language Package, which supports open source code; allows you to invoke analytic actions from Python, R, Java, JavaScript, and LUA).

Text Analytics

Get insights from texts thanks to the combination of natural language processing, machine learning and language rules.

The Text Analytics function allows you to build models based on training documents or procedures or other that analyze and categorize a series of documents, to optimize the value of your text-based data as well as categorize and extract the concept with a flow of visual programming and a linguistic methodology based on rules and grammars.

Parsing actions are provided in all 33 supported languages.

Model release, management and automation

Simplify the model lifecycle by recording, editing, tracking, evaluating, publishing and reporting on analytical models.

The following activities can be carried out:

Use a single interface to access all your designs and templates
Import models developed using a SAS application, an open source programming language such as Python or R, and PMML models
Compare models to evaluate candidate models.
Manage the data used for scoring, validation of published models, and performance monitoring
Perform scoring tests to validate models
Publish SAS models and open source models to supported publishing destinations
Automatically run monitoring processes to measure model performance on historical and current data. Performance benchmarking reports show the sample model's scoring performance and document compliance with required standards
Evaluate the decay of the model
Use native Python code in Jupyter notebooks to import, export and manage models

A centralized and searchable repository offers complete visibility into the analytical process and ensures traceability and governance. Lifecycle management and version control simplify model management, allowing you to track every step of the management process.

It is obviously possible to continuously update the models to reflect changing market and business conditions. You can retrain the existing model on new data or revise the model completely.

Quality check

Improve products and optimize processes thanks to statistical process control. Go beyond basic process control to incorporate more advanced statistical analysis for more insights into processes and product improvement.

You can monitor multiple processes and integrate a wide variety of data. Quality control allows you to maintain consistent standards and use all the information gathered to make better decisions.

There are basic tools for solving quality problems such as Pareto charts but also more sophisticated tools in the field of statistical process control: Shewhart chart or control Chart, trend chart for time dependent data, modeling and monitoring of processes multivariate and many others.

Optimization

Use a powerful range of optimization, simulation and project planning techniques to identify actions that will perform best, operating within resource constraints and other relevant restrictions.

Leverage an in-memory distributed engine to deliver optimization modeling results at blazing speeds.

Analysts, data scientists, and other optimization professionals can identify the actions that will produce the best results.

Data Preparation, a fundamental step

Data preparation is always a fundamental and necessary step to clean, integrate, validate the data inputs for statistical modeling but also for any other analytical process.

With SAS Studio, you can view your programs as a process flow:

Data Visualization with SAS Visual Analytics

Thanks to SAS Visual Analytics, users of all levels can visually explore the data.

SAS Visual Analytics provides with a modern and integrated environment for discovery and governed exploration of data. Even without advanced analytical skills, it is possible to examine and understand patterns, trends and relationships in data.

It’s easy to create and share reports and dashboards that monitor business performance.

Easy-to-use analytics and visualizations all help you gain insights from your data to better solve complex business problems.

It is the right solution for all those organizations that want to create, share and collaborate on insights. This includes decision makers, analysts, report creators and users of various kinds.

It also provides IT with an easy way to govern and manage data integrity and security.

SAS Visual Analytics comprises these key components:

Transforming elementary data into meaningful information

The evolution of production, management and interaction systems, also enabled by digitization, makes it possible to have large quantities of data available which, if not treated adequately and normalized, produce a lot of disrupted information, difficult to process and of little significance for the business strategy of any sector.

Blue BI realizes Business Intelligence solutions based on SAS systems capable of extracting significant information starting from elementary data, even uneven sources, in a short time (real-time) to obtain insights with a significant positive impact on all the decision-making processes of the company.

SAS solutions for every technology

In Blue BI we solve the problem of integrations and data normalization between different technologies (SAP, Qlik, Microsoft, SAS, Oracle, Tableau, Open Source and more), software already in use in the company and external sources.

We create and manage Business Intelligence & Analytics solutions that meet the needs of customers in terms of usability, reliability and achievement of the expected results and benefits, introducing the best technological solution in company organizations based on context and needs, without any conditioning by the Software Vendor.

Benchmark Blue BI

We keep ourselves constantly updated on all the technologies present on the Business Analysis market. One of the value-added services we offer is the drafting of Comparative benchmarks between the different BA solutions, built on the basis of customer needs and based on project operational experiences, in addition to the evaluation of product technical specifications.

Benchmark analysis is essential to support the customer in selecting the technology that best suits his needs.