In-Database Analytics: DB Lytix™
DB Lytix is Fuzzy Logix’s flagship product. It is a comprehensive library of over 800 mathematical and statistical functions used widely in data mining and analytics applications, including “big data analytics”. The introduction of DB Lytix has created a paradigm shift in analytics performance by introducing the concept of in-database analytics.
Traditionally, when running analytics, data is extracted from the Enterprise Data Warehouse (EDW) into a dedicated analytics server; the analytics engine (SAS/SPSS/etc) runs in this server and generates the results that are then served to visualization/reporting tools like Excel, Business Objects, MicroStrategy, Tableau. LogiXML etc. for downstream consumption. This process leads to the following challenges:
- The data extraction process from the EDW to the analytics server takes longer and longer as data volumes grow; this step can consume 60-70% of the total analytics processing time.
- The cost of maintaining the analytics infrastructure adds to the total cost of ownership due to hardware, software and staff expenses.
- The volume of data that we can run the analytics on is limited by the volume that can be extracted and stored in the analytics server. This means analysts are constantly working with smaller data sets than they would like and spend a great deal of time moving and managing data as opposed to analyzing data.
- Analytic results are not current or real-time. They are only as current as the last data extract from the EDW, which could be days or weeks old.
- Business users can only receive, but not run analytics since they can’t change variables and run the models. For example, if running segmentation by different categories, the marketing user is dependent upon the quantitative analyst to implement changes and rerun the models.
Using DB Lytix solves all of these problems by eliminating the data extraction, middle-tier analytics server and redundant data storage while allowing you to embed analytics into existing reporting tools so end users can run analytics on demand.
Using DB Lytix can produce high return on investment. A few examples of the value realized by our clients:
- 10% reduction in churn worth $5 million
- 3,200% return on investment in 12 months.
- 50% reduction in cost of sale on the Internet worth $3 million
- 8% increase in sales worth $3.2 million
- 1,000% increase in revenue
- 10X improvement in model development time worth $2.8 million
- 50% reduction in the cost of analytics worth $600,000
DB Lytix is used in critical areas by many Fortune 500 companies. We will gladly put you in touch with our customers who will vouch for our capabilities and value. We strive for, and have achieved 100% positive references. Example uses:
- Healthcare – provider scoring, preventative healthcare, pharmaceutical prescribing habit management, peer comparison and adverse drug ingredient reaction analysis.
- Banking and Financial Services – valuation of cash flows, equity derivative modeling, fixed income, time-series modeling and credit and market risk management.
- Digital Media –ad optimization, behavioral segmentation, forecasting and marketing effectiveness.
- Retail and Marketing – market basket, churn, recommendation engines (next likely purchase), marketing optimization and promotional effectiveness.
DB Lytix is running in production with terabytes and petabytes of data. Because we minimize data movement and have built our models to take advantage of data and processing parallelism, we offer the fastest and most scalable models available. A few benchmarks:
Since DB Lytix functions can be called using SQL, you can use almost any existing analytics or reporting tool to start experiencing the benefits today. Our customers use DB Lytix with MicroStrategy, Tableau, Excel, Business Objects, Flash/Flex, Cognos and other reporting and visualization tools. You can leverage your current investment in these tools while adding robust analytic capabilities and enhancing performance. For example one of our MicroStrategy customers saw a 91% improvement in performance. This feature makes it easy to deploy analytics across your organization with minimal disruption and maximum impact. For users of SAS and SPSS, you can call in-database functions directly from those programs and achieve tremendous gains in performance. For example, using SAS, you can write a PROC SQL statement and run DB Lytix analytics directly in the database. Many of our customers do this and see dramatic performance gains. A common use is to improve long running SAS processes that require a great deal of data movement and processing, customer scoring being one classic example. One of our customers saw their scoring time drop from 10 hours to 10 minutes by processing the score in-database with DB Lytix and moving only the resulting score back to SAS for future processing. If you have a large amount of existing SAS code, selective replacement of code will provide performance enhancement while allowing you to stay in SAS as your primary analytics tool.
|The Importance of Analytics for Business||In-Database Analytics Overview||Demos of Fuzzy Logix DB Lytix™|
Ease of Adoption
Installation: DB Lytix comes as an add-on package to your database and installs in 20 minutes; hence it is one of the most easily deployable and adoptable software products in the industry. Unlike traditional solutions, there is no additional hardware, datacenter space, duplicate storage or networking setup required.
Time-to-Action: The very paradigm of DB Lytix helps you gain efficiency in analysis, saves time and cost, and enables you to provide answers to the decision makers in minutes. Analytics can be built and incorporated into business reporting tools and processes in a few days or weeks.
Affordable: DB Lytix is available as a perpetual use license for the database platform where it is installed. There are no seat licenses, no user licenses and no limitations to the amount of data being analyzed or number of applications being serviced. This license methodology makes DB Lytix the most affordable high-performance analytics solution available and represents a welcome change from licensing practices of other providers.
DB Lytix is available on many industry-leading data warehouse platforms including Teradata, Aster Data, Netezza, ParAccel, Sybase IQ, SQL Server, MySQL, and Informix. In each implementation, our engineering team worked with the specific database product developers to ensure that our analytics engine leverages the best design attributes available in each platform. DB Lytix can be purchased from ParAccel, SAP, Quest Software (Dell) and other partners or directly from Fuzzy Logix. DB Lytix™ The DB Lytix library consists of statistical and data-mining functions written in C/C++ that have been embedded within the kernel of the Data Warehouse and can be invoked from simple SQL queries. Just as you can call native database functions like SUM, AVG, COUNT, you can now run segmentation, regression and 800 other models. A high level list of available models is on the back of our brochure. If you are interested in modelling in the financial services industry, FIN Lytix contains a comprehensive financial library that includes equity, fixed income, risk and time-series models.
Who Can Benefit From Using DB Lytix?
Decision makers need the ability to run analytics on demand – an impossible task with traditional solutions. DB Lytix allows models to be quickly integrated into existing software that business users can run at will. For example, the ability to run segmentation, forecasting or marketing optimization using a variety of input values will allow business users to quickly gain the insight they need.
Data Scientists and Quantitative Analysts
By eliminating the time required to transport data from storage to analytic servers and running DB Lytix models that take advantage of data and processing parallelism , our customers tell us they can build and execute 10 times as many models than they can with traditional analytics software. The library can be used to replicate existing solutions and create new ones with dramatically improved performance. Many of our customers have also eliminated sampling, coefficient reduction and other time consuming tasks because the performance and scalability of the models eliminates the need for these tasks. DB Lytix allows quantitative analyst to analyze enormous amounts of data and iterate through the model discovery process faster than anything else on the market today.
Software and Application Developers
Embedding the analytical functions in DB Lytix into software products is seamless and quick. By adding analytical functions such as forecasting, churn, recommendation engines and behavioral segmentation to existing software products, applications, and business processes companies can differentiate their products and create value for customers.