Patent Application
20100175019
Business intelligence (BI) software is currently a $20 billion market. The goal of BI software tools is to provide historical and analytical data for aspects of a business including sales, marketing, management reporting, business process management, budgeting, forecasting, financial reporting and similar areas. Currently, sophisticated tools exist for the creation of reports and graphical interfaces that present predefined views of business data. For example, prior art
Conventional dashboards such as that shown in
This limitation in current BI software is more unfortunate for the fact that current multidimensional databases are constructed so as to be able to provide a very broad range of information and comparisons relating to all aspects of a business. Current multidimensional databases are organized into multidimensional cubes, which consist of numeric data, referred to herein as measures, which are categorized and defined by a variety of characteristics, referred to herein as dimensions. Each measure may be viewed as being the result of multiple dimensions. For example, a company might wish to analyze some financial data by product, time-period, city, type of revenue and cost. Each of these factors is a dimension which, taken together, determine the financial data measure.
Each of the elements in a multidimensional cube database may also be organized into a hierarchy. The hierarchy is a series of parent-child relationships, typically where descendant dimensions are subcategories of a parent dimension. The parent dimension may itself be one of many sub-categories of a grandparent dimension, and so on. As examples, cities may be subcategories of a region, which is in turn a subcategory of a country; products could be summarized into larger categories; and cost headings could be grouped into types of expenditures, etc. Conversely, it is possible to start at a highly summarized level, and drill down into the cube to discover descendant subcategories. Dimensions within a given category or subcategory are referred to herein as siblings of each other, and are cross-referenced to each other within the multidimensional cube database.
Given the vast amount of cross-referenced vertical and horizontal data within a multidimensional database, it is desirable to provide a BI tool which escapes the paradigm of algorithms that are written to convey only specific aspects of a business's historical and analytical data.
Embodiments of the present system in general relate to an ad hoc business data exploration tool providing guided access to the vast amount of data within a multidimensional database. The tool guides the user by suggesting insights which may be of particular interest to the user based on a scoring of the insights and user feedback on desirable/undesirable insights. The present system works in conjunction with custom algorithms, referred to herein as reusable business logic algorithms, as well as a conventional multidimensional database.
When viewing a report, a user is able to select a given measure, and launch the business exploration tool, also referred to herein as an “insights” tool, to learn more detail about the selected measure. Upon launch, the user is presented with a dashboard including windows having an appearance of typical reports. However, the dashboard includes certain fixed and variable insights into the selected measure. A fixed insight is one that is automatically presented to the user regardless of the measure selected for further analysis. Alternative embodiments of the present system may operate without fixed insights. A variable insight is one that is selected by the present system for display to the user. In particular, using heuristic rules, the present system selects what appear to be the most interesting insights into the selected measure from a large number of stored insights. The present system may display different numbers of variable insights to the user in different embodiments.
A further aspect of the present system allows users to view additional dashboards focusing on sibling dimensions of the dimensions used to formulate the selected measure. In embodiments, these sibling dashboards may be displayed to the side of the original insight dashboard. The present system also allows users to drill down into a selected dimension to provide insights into descendants of the selected dimension. The insights which are displayed are those which appear to be the most interesting insights based on the selected dimension. In embodiments, these descendant dashboards may be displayed below the original insight dashboard. By interacting with the present system in this manner, a user may navigate to a variety of different dashboards, each including insights selected by the present system as being the most interesting. In this way, a user may access the full power of the multidimensional database by discovering worthwhile information the user may not have otherwise found or been interested in.
Embodiments of the present system will now be described with reference to
While the BI server 100 is described below as a single machine, it is understood that the below described components of BI server 100 may alternatively be distributed across more than one machine. For example, it is understood that a first server may have the BI algorithms according to the present system and a second server may be a separate web server. Moreover, the multidimensional database, described below, may be incorporated into BI server 100 in further embodiments.
In an embodiment where the BI server 100 comprises a single machine, the BI server may include one or more processors 104, as well as an operating system 106 and one or more program applications 110 executed on processor 104. The application programs include an insights tool application program and the reusable business logic algorithms, both described hereinafter. System memory 116 may further be provided for use by processor 104. The memory 116 can be implemented as a combination of read/write memory, such as static random access memory (SRAM), and read-only memory, such as electrically programmable read only memory (EPROM). A network interface 118 may also be provided to enable communication between the BI server 100 and computing system 102. The BI server further includes an insight scoring engine 114 for ranking insights and determining which insights to present to users as described hereinafter.
The BI server 100 communicates with a multidimensional database 120. A variety of multidimensional databases 120 are known for use with the present system, including for example Microsoft SQL Server Analysis Services, Hyperion Essbase, IBM DB2 OLAP, and SAP BW. Others are contemplated. As is known, such databases organize data into multidimensional cubes, which consist of numerical measures defined by a number of dimensions.
As is further known, the dimensions in database 120 may have a hierarchical tree structure of categories and subcategories, with dimensions in the same category referred to as siblings of each other. Moreover, one or more of these dimensions may have a subcategory of descendant dimensions, with each descendant dimension in the subcategory being siblings of each other, and so on. The dimensions from different categories/subcategories may be used to derive a numerical measure. Known application programming interfaces (APIs) may be used to allow communication between the multidimensional database and the reusable business logic algorithms to allow extraction of measures and dimensions from the database for presentation in accordance with the present system as described below.
The operation of the present system will now be described initially with reference to the report shown in
Upon launching the insights tool in step 200, the processor presents an insights dashboard 140 over the display as shown for example in
In addition to window 144, the dashboard 140 may include one or more fixed insight windows (150 and 152 in
The insights presented in the fixed and variable windows are generated by reusable business logic algorithms. These algorithms may be created, for example by an IT administrator for a business, and stored for use by the insight tool of the present system. Where these algorithms were used in the past as dedicated code for creating a specific report, the business logic algorithms used in the present system are said to be reusable in that they are generalized to accept different inputs so as to provide some history, analysis, comparison, forecasting, etc. relating to any selected measure. Thus, where a user selects a first measure, a particular business logic algorithm may be used to provide insight and detail with regard to that measure. If the user then selects a second measure, the same business logic algorithm may again be used to this time provide insight and detail with regard to the second measure.
In the example of
Depending on which tab is selected, the fixed insight window 152 shows a graph illustrating the dimension used in forming the measure of window 144 together with a comparison against its siblings. As explained below, a user can select a tab, and then select a sibling from the graph, and the present system will present another dashboard along the side of dashboard 140 giving a more detailed analysis of the selected sibling.
While
The insights dashboard 140 of
In the embodiment of
It is understood that any of the measures displayed on the report of
Insight scoring engine 114 determines which insights would appear to have the most interesting features and/or trends to display in the variable insight windows on dashboard 140. In particular, a large number of insights may be generated from different business logic algorithms, and each of the insights may be stored. Using a heuristic approach, the scoring engine 114 ranks each of the insights, and presents the top insights to the user in the variable insight windows as described above. A number of heuristic rules may be applied in selecting the best insights.
In embodiments, such heuristic rules may perform a top to bottom exhaustive search over each of the dimensions in the database to determine which dimensions most significantly affect the overall measure. The insight scoring engine 114 may then select the insights which best show the effects those identified dimensions have on the overall measure. The dimensions examined may be one or more of the dimensions directly defining the overall measure, or a sibling or subcategory of a dimension. Dimensions which are further away from directly impacting the overall measure may have a lesser likelihood of being selected as a top insight. However, even dimensions which are remote from the overall measure may generate a top insight if such dimensions have a large impact on the overall measure.
There are a variety of heuristic rules which may be used to select interesting insights. One such heuristic rule looks at key contributors to a selected measure. The present system searches for a few dimensions or combinations of dimensions that have a disproportional contribution to the overall pattern. This analysis may be performed a number of ways, but in one embodiment, the present invention examines all dimensions which relate to a measure, and sorts all dimensions in a given category in increasing order. For example, the present system may take all months which contributed to a measure, and sort them in ascending order of contribution.
The present system may then examine the skewness of the dimensions in the category. For a given category of Y1, Y2, . . . YN dimensions, the skewness of the data may be determined. As is known, skewness is a measure of the asymmetry of the distribution of the dimensional data for a give category of dimensions, and is given by:
where
If the skewness is very negative (below a certain threshold, e.g., −2) then a pattern of key destructors exists, which may also be an important insight. The selection of elements, or dimensions, is done in a similar way as in positive skewness (i.e., key contributors). If the skewness is between the two thresholds, e.g., bigger than −2 and smaller than 2, then no significant pattern is observed.
Another heuristic rule which may be used to find insights is an examination of trends in the dimensions bearing on a measure to find a difference in a pattern. Here, the objective is not to check whether key contributors exist, but rather to analyze if the pattern has changed along a certain dimension. An example is to look sales of a product along a time series, for example this year and last year, and to check whether the sales in a certain month varies significantly and unexpectedly. For example, sales may rise during the holiday months, but did they rise above or below expectation. This analysis may be performed for example using a known Chi-square test.
Once differences in a series are identified by the Chi-square test, the present system examines whether the differences are expected. Unexpected trends are identified by normalizing the dimensions used in the identified trend, and looking at the skewness of the normalized series. Major and unexpected contributors may then be identified using the key contributors heuristic discussed above.
Once insights have been developed according to the above and other heuristic rules, the insights are prioritized. There may be several prioritization methods: by statistical significance, by importance, by experience, by “wisdom of the crowd”, by similarity, etc. Each of these is set forth below.
In prioritization by statistical significance, each algorithm should come with a significance value of its results (e.g., the significance of the chi-square test). The algorithms are prioritized then by that unbiased measurement.
In prioritization by importance, the algorithms or the data they look at each are assigned a different business importance. The importance can be manually defined by the user, predefined by the system or measured by some other method. The observations can then be presented and ranked according to this consideration.
In prioritization by experience, the system presents insights that are similar or relevant to insights that the user has already seen and studied (because these prior insights are assumed to be relevant and of interest to the user. A complementary method is to exclude (give a lower priority) observations that were previously studied (under the assumption that the user wants to learn something new).
Prioritization by “wisdom of the crowd,” is a collaborative filtering strategy that presents observations that were viewed by other people, preferably people who are similar to the current user (holding similar position, geography or more).
Prioritization by similarity presents observations that are similar (or contrary, very dissimilar) to a specific insight.
Still further methods of prioritizing insights are contemplated, including those accomplished manually or by machine learning and artificial intelligence.
As an alternative to an online single dimension exhaustive search, the scoring engine 114 may use offline data preprocessing. The heuristic rules may further take into account user profile and feedback. In particular, as explained below, users are given the option of providing feedback to rate the insights which are selected for display.
In step 210, the insights tool checks whether the user has moved the pointing device to hover over a particular insight. If so, the insight tool presents more detailed information relating to that insight in step 212. For example,
Instead of hovering over an insight in step 210, a user may instead select a dimension in one of the above-described windows 144, 150, 152, 156 and 158 in a step 214. If the user selects a dimension from the graph within the sibling relation window 152 in step 230 (
In embodiments, only one sibling dashboard 170 may be displayed to the side of the original insights dashboard 140. This sibling dashboard 170 may be displayed to the right of insights dashboard 140, except in a situation where the category being examined is time. In such an example, if an earlier time period is selected for presentation in a new dashboard 170, that dashboard 170 may be displayed to the left of insights dashboard 140. If a later time period is selected for presentation in a new dashboard 170, that dashboard 170 may be displayed to the right of insights dashboard 140.
In further embodiments, the sibling dashboard 170 may include a sibling relations insight window 172 similar to sibling relations insight window 152 shown and described above. In such embodiments, a user may select a particular sibling from sibling relation insight window 172 (either from the same category of siblings selected from window 152 or from a different category). This selection may result in a third dashboard (not shown) opening up to the side of dashboard 170, showing the selected sibling dimension, the resulting measure, and fixed and variable insight windows. In such an embodiment, any number of horizontally oriented dashboards may be displayed. As the display may not be large enough to display all dashboards in a horizontal row, a scroll bar may be presented in a known manner to allow a user to scroll through the various dashboards in a row.
Instead of selecting a sibling from a category of sibling dimensions in relation windows 152, 172 in step 230, a user may instead select to drill down into a descendant subcategory of a dimension in step 236. Dimensions having further descendants may be indicated as hyperlinks on dashboard 140 in
Dashboard 180 shows a window 184, similar to window 144 of dashboard 140, including a net revenue in the service advertiser industry attributed to that specific service advertiser in the U.S. in 2008 of $396,741.88. Dashboard 180 may further include the fixed and variable insight windows described above. For the variable insight windows, the present system may select insights that appear to be of greatest interest to the user with respect to details of the parent dimension selected from dashboard 140. For example, upon drilling down into a dimension, the present system may show insights relating to dimensions that contributed most significantly to the parent dimension, or relating to trends or counter-trends in the descendants. Other insights relating to descendants are contemplated.
In embodiments, there may only be two vertically oriented dashboards. In further embodiments, the dashboard 180 may have hyperlinks allowing a user to drilldown further into subcategories of the dimensions shown in dashboard 180. A user may also have the option of selecting a sibling from the sibling relations window 192 in dashboard 180, so as to open up one or more additional dashboards vertically below the dashboard 180.
In the above-described manner, a user may generate a detailed map of horizontal siblings and vertical drilldown detail relating to one or more measures and dimensions of that measure. By interacting with the present system in this manner, a user may navigate to a variety of different dashboards, each including insights selected by the present system as being the most interesting. In this way, a user may access the full power of the multidimensional database by discovering worthwhile information the user may not have otherwise found or been interested in.
Once a user opens up a new dashboard to the side or below the insights dashboard 140, that new dashboard may be displayed more prominently than the other dashboards. For example, the newly displayed dashboard may be larger, where the other dashboards may be smaller and have a degree of transparency. A user may move to another dashboard, as by hovering over it, to make that dashboard larger and not transparent. A user may elect to exit the drilldown/sibling map of
In step 216, the insight tool looks for user feedback on the insights which have been selected as the best insights for display in the variable insight windows on the dashboard 140. Any feedback received in step 216 is stored, for example in an XML file, and provided to the scoring engine 114 so that the scoring engine can use that feedback when evaluating all of the stored insights. Although not shown in the figures, each window displaying a variable insight may include a feedback object 154 that allows a user to indicate whether they would like to see, or not see, that particular insight in the future. In embodiments, the feedback object may be a “thumbs-up/thumbs down” indicator allowing users to indicate their approval or disapproval, respectively of a given insight. In embodiments, once a user gives an insight a thumbs-down, that insight may not be shown to that user in the future. Conversely, if a user gives an insight a thumbs-up, that insight may be presented to the user in the future until a user changes the feedback for that insight. Over time, the scoring engine hones the insights which are selected for display to better reflect the insights the user would most like to be shown.
In further embodiments, in addition to receiving feedback on insights, the present system may make recommendations on insights. If a user provides positive feedback on a given insight, a pop-up window may be presented including a statement along the lines of, “users who liked this insight also liked . . . . ” The system may then recommend one or more additional stored insights, which may be presented as hyperlinks for selection by the user. The present system may identify correlations between insights in a known manner.
In step 218, the user further has the option to exit the insight tool. If the user elects to exit the insight tool, the user is returned to the report shown in
Following is one example of how the insight tool of the present invention may be used to discover additional and useful information about an aspect of a business from the vast amount of data collected and stored in the business's multidimensional database.
In this example, a user is a sales manager for a food company. He dedicates an hour at the end of every week to learn about the well-being of his region. He starts by opening a report which shows the top five and bottom five stores in his region. He sees that Store #5 is up by 23% compared to the previous week. He remembers that there was a big marketing effort for fine wine in the region. He would like to know more about the region and the stores in it. He drags the figure to the insights button to launch the insights tool.
At this point the system builds the story behind the sales of Store #5 for the previous week. It automatically compares the sales to previous weeks and looks for highlights, lowlights and other interesting facts. In addition, it creates links to other stores, sub regions, products, customer segments, etc.
The dashboard displays an informative but comprehensive picture of the top scored insights. The UX is familiar to the user, as it looks like other reports that the user's IT administrator generated in the past. The user finds one of the facts not so related to the task at hand and is able to provide a “thumbs down” feedback. Other facts are very interesting and the user gives them a “thumbs up”. After doing so, the user learns that even though the marking effort was for fine wine, sales for beers and other alcoholic beverages are up by 35%. Based on this learned and unexpected information, the user decides to widen the marketing campaign.
The present system is operational with numerous other general purpose or special purpose computing systems, environments or configurations. Examples of well known computing systems, environments and/or configurations that may be suitable for use with the present system include, but are not limited to, personal computers, server computers, multiprocessor systems, microprocessor-based systems, network PCs, minicomputers, hand-held computing devices, mainframe computers, and other distributed computing environments that include any of the above systems or devices, and the like.
The present system may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform particular tasks or implement particular abstract data types. In the distributed and parallel processing cluster of computing systems used to implement the present system, tasks are performed by remote processing devices that are linked through a communication network. In such a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
With reference to
Computer 100 typically includes a variety of computer readable media. Computer readable media can be any available media that can be accessed by computer 100 and includes both volatile and nonvolatile media, removable and non-removable media. By way of example, and not limitation, computer readable media may comprise computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVDs) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by computer 100. Communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared and other wireless media. Combinations of any of the above are also included within the scope of computer readable media.
The system memory 116 includes computer storage media in the form of volatile and/or nonvolatile memory such as read only memory (ROM) 331 and random access memory (RAM) 332. A basic input/output system (BIOS) 333, containing the basic routines that help to transfer information between elements within computer 100, such as during start-up, is typically stored in ROM 331. RAM 332 typically contains data and/or program modules that are immediately accessible to and/or presently being operated on by processing unit 104. By way of example, and not limitation,
The computer 100 may also include other removable/non-removable, volatile/nonvolatile computer storage media. By way of example only,
The drives and their associated computer storage media discussed above and illustrated in
A user may enter commands and information into the computer 100 through input devices such as a keyboard 362 and pointing device 361, commonly referred to as a mouse, trackball or touch pad. Other input devices (not shown) may be included. These and other input devices are often connected to the processing unit 104 through a user input interface 360 that is coupled to the system bus 321, but may be connected by other interface and bus structures, such as a parallel port, game port or a universal serial bus (USB). A monitor 391 or other type of display device is also connected to the system bus 321 via an interface, such as a video interface 390. In addition to the monitor 391, computers may also include other peripheral output devices such as speakers 397 and printer 396, which may be connected through an output peripheral interface 395.
As indicated above, the computer 100 may operate in a networked environment using logical connections to one or more remote computers in the cluster, such as a remote computer 380. The remote computer 380 may be a personal computer, a server, a router, a network PC, a peer device or other common network node, and typically includes many or all of the elements described above relative to the computer 100, although only a memory storage device 381 has been illustrated in
When used in a LAN networking environment, the computer 100 is connected to the LAN 371 through a network interface or adapter 118. When used in a WAN networking environment, the computer 100 typically includes a modem 372 or other means for establishing communication over the WAN 373, such as the Internet. The modem 372, which may be internal or external, may be connected to the system bus 321 via the user input interface 360, or other appropriate mechanism. In a networked environment, program modules depicted relative to the computer 100, or portions thereof, may be stored in the remote memory storage device. By way of example, and not limitation,
The foregoing detailed description of the inventive system has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the inventive system to the precise form disclosed. Many modifications and variations are possible in light of the above teaching. The described embodiments were chosen in order to best explain the principles of the inventive system and its practical application to thereby enable others skilled in the art to best utilize the inventive system in various embodiments and with various modifications as are suited to the particular use contemplated. It is intended that the scope of the inventive system be defined by the claims appended hereto.
