The "Contour Standard" 3.0 is a personal tool for corporate data analysis. It performs the following tasks:
The program supports decision making, analysis and reporting using data arrays, accumulated in automated banking and accounting systems and/or other OLTP modules and separate databases, without programming and administrating.
- Querying databases of any type;
- Performing operational data analysis;
- Generating presentation-quality analytical reports with diagrams and graphs "on snap";
- Publishing reports on the Web.
Comparing with the previous version the "Contour Standard" 3.0 includes the company's brightest new feature - the Contour Cube OLAP technology component.
Lying in the heart of the Contour family products, the new Contour Cube component performs different kinds of complicated multidimensional analysis (including statistic and cluster ones) on data consisting of hundreds of thousands records. Dataset of under 400,000 unique records can be fairly processed by a PC with 64MB RAM, and of under 1,000,000 records by a PC with 128MB RAM (the data is for an instance table of 7 dimensions and 2 measures). Thus Contour Cube gives personal systems power and capacity which were formerly only available for powerful OLAP servers.
The Contour Cube
The Contour Cube is a next generation, high-performance, multi-functional ROLAP/DOLAP component. It has layered architecture consisting of the OLAP engine, data access interface for getting data from different sources, and GUI.
The key features of the component are:
Processing of huge amounts of data
Tests on PC with Intel Celeron 400 processor and RAM 64 MB shows that average time of rotating the cube with 7 dimensions and 2 measures derived from table of 60,000 unique records, is 0.6 seconds. Average time rotation time for the cube derived from a table of 400,000 records on the same computer is less than 7 seconds on average and less than 10 seconds in the worst case!
These are the best results among the current OLAP components that are known to us. The best of counterparts, Data Dynamics'
DynamiCube, has characteristics that are 2 to 10 times worse on moderate data sets (over 50,000 records) and 5 to 50 times worse on large amounts of data (over 250,000 records).
The Contour Cube blends features of the best OLAP components, but also includes some unique features, which don't present in any currently existing components.
The common features are:
- Multiple measures (more than one measure dimension per cube).
- Multiple filters on dimensions.
- Generating both standard and custom time intervals ("Year", "Quarter", "Month", "Week", etc.) on "Date" type dimension.
- One dimension value (branch) opening/closing.
- One dimension compression/uncompression (drill up/drill down).
- Automatic diagram control (graphs are "live", i.e. automatically follow cube's rotations).
- Manual diagram adjustment.
- Zero columns/rows filtering, to compress sparse tables.
- Swap rows and columns (transposition). At that, table columns and rows swaps (i.e. columns becomes rows and vice versa). Applied to improve analyst's comprehension of data and to select the best printing view.
- Export to HTML and .xls (Excel) formats. To analyze data with MS Excel, create a free format report and publish report on the Web.
- Show values as percentage of row or column totals.
The unique features are:
- The "Account balance" aggregating algorithm. Due to the main purporse of OLAP components to be tools for "accumulating" (summing) types of analysis, e.g. sells analysis, budget analysis, etc., they use the same approach on account balances. This is not right, because a quarter account balance is not the sum of everyday account balances, but is the balance of the last day in the quarter. Implementation of this algorithm allows a user to apply the component for balance analysis and makes it useful not only for economists and marketing specialists, but for accountants as well.
- View statistical characteristics of values "oh the fly". For every shown measure there can be shown some "virtual" measures designating statistical values for that measure's values (subtotals). Such virtual measures can be created "on-the-fly" (i.e. for already displayed slice) with two mouse clicks. The statistical values include Count, Average, Average Deviation, Variance, Standard Deviation, Variance Ratio and Root Mean Square.
Minimal memory requirements
The Contour Cube requires less RAM for data processing (7 MB per table of 60,000 records with 7 dimensions and 2 measures).
These outstanding characteristics have been achieved via unique mathematical model created by the company specialists. The component architecture is multi-layer. The OLAP Engine is the base layer of the component. It's implemented as cross-platform library that provides abstract API for use by different visualization layers. This API has data loading functionality, multidimensional cube slices computing capability, analytical and service functions. The OLAP Engine layer itself consists of calculating machine and abstract multidimensional data warehouse. The latter allows saving cube's data in a file in a very compact form (the "microcube") so it can be transmitted to other users or analysed in disconnected mode (without database).