SAS Consulting

Summary

Due to specialization, the majority of IT projects currently require many consultants and project management resources. At BeOptimized, our consultants are multitasking: Data Manipulation, Analytics as well as Business Intelligence. Our certified experts have been using not one, but all of SAS software for many years. Thanks to this versatility, our project management costs remain minimal. Our developments are optimized and our projects are delivered on time.

Want to get an example of what we did for a customer? Download here the SAS Forum 2016 presentation to see part of the tasks we did at ENGIE Electrabel.

Quick Facts

  • Services delivered by experienced ex-SAS consultants who worked 8 years at SAS Institute
  • Our experts are SAS certified with excellent business understanding
  • End-to-end approach
  • 100% adapted to your needs
  • Possibility to be coached on the development when the project is done

Certifications

Service Description

ETL jobs (Extract, Transform & Load) are the starting point of every projects. They establish a link between the data and the process your are working on. At BeOptimized, we acquired an excellent expertise in these jobs (in Base SAS, as well as in Data Integration Studio). We are using them since many years. This include:
  • Creation of SAS extractors to convert data into SAS from various systems (CSV, Databases Tables, XML...)
  • Creation of Data quality rules in phase with process needs
  • Fuzzy match code creation using DataFlux algorithms
  • Data combination using in-database, in-memory or more traditional technics
  • Automatic creation of inconsistency reports & email sending
  • Implementation of Slowly Changing Dimension
  • Load/update data in an efficient way
  • Security
  • Impact/reverse impact analysis
  • Scheduling
  • Log analysis for performance improvement
  • Creation of .spk packages and promotion to other environments
With BeOptimized we are able to discover and predict unknown patterns in your large data sets. Our DataMining projects are generally delivered using the following methodology:
  • Business understanding analysis
  • Analytical Base Table (ABT) creation
  • Utilization of the SEMMA Methodology to create and select the best analytical (supervised and unsupervised) model
    • Oversampling and splitting of dataset
    • Outliers & missing values management, transformation
    • Model generation (regression, decision tree, neural network...) and comparison using ROC & Lift curve
  • Scoring code Generation
  • Possibility to convert the datamining code into web service/ Stored Process
  • Model recalibration to today's situation
Thanks to the SAS BI Framework and our expertise in those tools, we deliver end-to-end reports which best reflect the data you are working on. From static printable report to interactive web-based reports, we master the entire SAS V9 BI platform.
  • Creation of reports using SAS BI Framework (VA, Stored Process, AMO,....)
  • Batch reports/ Real-time reports creation
  • Dynamic web based/MS office reports creation
  • Utilization of ODS, proc Graph, Gplot, Template.... and other proc misknown by SAS community
  • Security
  • Scheduling
  • Adaptation of report to today's situation
After using your SAS processes for several years, you may have noticed a significant decrease in performance. Indeed, the data volume, number of concurrent processes and the number of people using your programs can significantly decrease your code/process performance, as they were created for a very specific framework a few years before. At BeOptimized, our experts have the knowledge that will allow you to monitor, analyze and improve your SAS codes using advanced programming techniques. Our high level planning is organized as follow:
  • Analysis of the end-to-end process to detect possible programming shortcuts
  • Log analysis to detect time consuming processes
  • Application of state-of-the-art SAS techniques (In Memory, In Database, In Parallel...) to reduce bottlenecks to an acceptable level
  • Actual/new output comparison to certify that the expecting result didn't change