 |
|
|
Archive - Edition 2: Reporting
Leading Article
Current Trends in Reporting
Content
- Creating Reporting Structures through Information Modeling
- Structured versus Unstructured Information
Information Visualization
- Orion – A Visual Approach to Business Analytics
- Knowledge Reporting with InfoZoom®
- Making the Truth Visible – Visual Design of Visual Reporting Tools
Selected Portals
Links & More
 |
Print version |
Related Links |
|
 |
Orion – A Visual Approach to Business Analytics |
|
Making the Truth Visible – Visual Design of Visual Reporting Tools |
Background Links |
|
|
|
Knowledge Reporting with InfoZoom®
Simply know more: Visual Data Mining and Lean Data Warehousing for decision makersBy Dr. Hans-Günter Lindner, CEO and Co-Founder human IT, Human Information Technologies – December 22, 2000
Disclaimer: Please note that this edition was written in 2000. Therefore, statements in the articles, particularly those regarding SAP's products, product strategy, branding strategy, and organizational structure, may no longer be valid.
Problems and a New Solution
Beautiful listings and business diagrams published in annual reports for shareholders reflect a common understanding of reports. It has only little to do with reporting, the process of developing such analyses, and the interests of the information customers, the shareholders as well as internal and external analysts. A report is only a bookmark for the manifestation of interest in information on a simple sheet of paper. Everybody knows the long way to knowledge, the missing individual focus and the one-way communication of reporting. Why is this the case? Because reports are mentally connected to paper but paper is a monologue. You read a report and you will have even more questions. Who answers? Specialists? Analysts? When? There is no interaction. Who cares about your action needs, your interests? Is this a goal or a problem?
Classical reporting tools do not fit the requirements of shareholder-oriented companies. Classical tools are only solutions for old problems and an outdated mental model from a time when reports were only paper documents. They lack of flexibility, intuitive usage, and interactivity. The pragmatic context of action and the users’ interests are absent.
Knowledge Reporting is the solution for current and future problems. It is a based on the technique of Knowledge Browsing which combines intuitive usage with Visual Data Mining and a Lean Data Warehouse. Visual Data Mining is the visual recognition of conspicuous patterns in data naturally done by humans for the extraction of knowledge and insights. A Lean Data Warehouse is a condensed database for decision support, reducing the size and costs of conventional Data Warehouses as well as their project duration. This leads to fast and efficient deployment of conventional Data Warehouses and an optimised design to fulfil the decision makers goals and interests. Business Intelligence in the context of Knowledge Browsing means to reduce the information overload to expose the entrepreneurial intelligence within decision makers.
The InfoZoom® Approach - Reducing the Information Overload
InfoZoom® is a Knowledge Browser that allows users to easily navigate through tables and databases. The user builds knowledge through interaction with InfoZoom®. Results can be bookmarked for dynamic information landscapes, called Infoscapes, or classical reports. Building knowledge is enabled by unique visualization, fast and interactive analysis, and intuitive Visual Data Mining.
InfoZoom® displays database tables with attributes as rows and objects as columns. In figure 1 each column corresponds to the sale and revenue of a textile order, characteristics of the order, time of sales and locale specifications. In our example, we added a classification rule separating good, medium, and low revenues using a standard traffic light metaphor; all attribute values are automatically colored according to this revenue-based classification. Most of the sales shown in figure 1 resulted in good revenue.
Figure 1: Textile stores' sales in a conventional table - click image for larger version
To gain an overview over the entire sales activities, you only have to click on the button Overview: all columns are compressed to fit a single window and are sorted independently of each other (Figure 2). The result is not a table layout but something like a set of stacked bar charts, each displaying the value distributions for one of the attributes. The distribution of the values show the entire sales database of over 9000 orders. The coloring provides insight in the quality of all groups of values. For example, "Hamburg" is the only store location that is colored green as are "shirts" as an article class. Especially "pilot shirts" but also "Tennis socks" are (on average) good revenue contributors. Merchandise is distributed similarly over all regions and store locations. "Nylon blouse" orders and orders of articles in size "x-small" are more frequent, but this results only in lower revenue per sale (red color).
Figure 2: Overview over all textile stores' sales - click image for larger version
To see all values correlated with each other we change into Compressed Mode. The column width is again reduced until all the objects fit on the screen (Figure 3). The width of about forty columns amounts to one pixel. In larger tables the (virtual) column width will be even smaller. Some techniques make the table readable despite this compression. The most important one is that neigh-boring cells with identical values are combined into one larger cell. The width of each cell indicates the number of subsequent objects with this value. If a cell is too small to display a numeric value, a point (horizontal line for wider cells) still indicates its relative height.
Figure 3: 9238 textile stores' sales sorted by revenue and office location - click image for larger version
Because the table was sorted by the attributes Revenue and Office, you can see the sorted revenue by office. As a result one immediately notices that Hamburg has not only higher total revenue; the quality of the sales activities is different. Hamburg has nearly the same distribution of sales activities like the other offices but the rate of low-valued sales is lower and the high-valued sales reach levels not matched by the other locations. This can be seen by comparing the patterns of the revenue curves: all office except Hamburg show a conventional hockey-stick effect and the maximum values are higher.
Instead of selecting attribute values from the values menu an attribute can also be restricted by selecting and double-clicking a value or value-range direct-ly in the table. In a short animation, clicked cells grow while the others shrink resulting in zooming into the table.
Figure 4: Zooming into Frankfurt and Wiesbaden sales on saturdays - click image for larger version
Like formula-cells in a spreadsheet program derived attributes can be defined which are automatically updated according to the user’s interests and context. For example, aggregated revenue will be recomputed after each zoom operation as can be seen in figure 5. In order to make two slices through the sales data, one per office and one per article class, one only has to drag and drop the two respective attributes Office and ArticleClass into the revenue cube.
Figure 5: Two slices inside the cube for revenue analysis - click image for larger version
As a result, we can see that the highest revenue was generated by dress sales in Hamburg and that in general Dresses are placed in the higher areas of revenue. Zooming into Dresses reduces the data to 2573 order records (Figure 6). The sums are now visualized by numbers as well as lines because the information density allows for the clear display of numbers. Additionally, we can see that high revenues correspond to high sales in Hamburg.
Figure 6: Zoom into Dresses based on figure 5 - click image for larger version
This resulting table can be transformed easily into a conventional report by simply clicking on the Report icon. Additional reports can be simply tailored by zooming in the Infoscape.
Business Examples
InfoZoom® can be used for the visualization of data across an unlimited number of domains. Actually, three areas of usage are mainly important since information overload in these areas is a mission-critical problem: e-Commerce, Business Intelligence, and System Management. An example for an e-Commerce application of InfoZoom® is a comparison of more than 4000 cars with more than 60 characteristics on the ADAC web-site (Figure 7). The ADAC automobile catalog is presented to end-users for an easy and intuitive comparison. The users’ feedback is highly encouraging: Easy high-speed access on the web for more transparency is now a reality.
Figure 7: Automobile catalog of ADAC - click image for larger version
The second example covers controlling and management information. The demo showcase Candy Corp, Inc. stands for several success stories for controlling and analysis of corporate information at Top 500 companies such as Deutsche Telekom, Nokia, Deutsche Bank as well as medium-sized enterprises such as Schuh Bär, Atelier Goldener Schnitt. SAP delivers a standard interface for InfoZoom®, the Knowledge Browser can be used plug & play for all SAP R/3 modules.
Figure 8: Corporate controlling system with InfoZoom® - click image for larger version
System management information is highly mission-critical. Several minutes for searching for root causes of problems and for detecting conspicuous patterns in big server systems that indicate problems are very expensive. Breakdown and lowering of IT-service quality can lead to immediate losses of sales and customers. To monitor big server systems, InfoZoom® is used to give insight into event messages from PATROL, a system management software from bmc software. As an SAP-outsourcer, AI Informatics monitors SAP R/3 to assure the quality of their systems. An example of InfoZoom® for PATROL is shown in figure 9.
Figure 9: Monitoring a server system with InfoZoom® for PATROL - click image for larger version
Vision for InfoZoom®
InfoZoom® assures easy access to complex information in real-time. Since several success stories show the clear benefit of InfoZoom® with low Total Cost of Ownership InfoZoom® makes its way as the standard tool for Knowledge Browsing in e-Commerce, Lean Data Warehousing and Visual Data Mining. Combined with humanIT’s Dynamic Personalization Server it allows an individualised information supply that is based on accurate, complete, and timely data. InfoZoom® leads to an intuitive and easy way to share knowledge in a way that is fun: Enjoy your data!
|
|
