Drilling and Information Analysis

Owned by Nicolás Palombo (Unlicensed)
Last updated: 08 Dec, 2008Version comment
3 min read
Unknown macro: {scrollbar}

Ideasoft O3 Browser potential at the time of analyzing information lies in the ease to perform queries and a quick navigation (drilling) from one view to the next one, turning the analysis activity from a reflexive activity on static data, into an interactive task, where the user proposes hypothesis, generates queries and obtains immediate answers that enable them to have a better understanding of any given situation.

The advantage of O3 Browser over other systems is that information is organized in structures or models that can be handled directly by the users through graphic commands and by means of the mouse and keyboard.

The structures or models used by O3 Browser are referred to as multidimensional as information on the different topics or events to be analyzed are simultaneously organized according to all the attributes or data that can be used to sort them out.

A multidimensional model or cube is the representation of a series of facts by means of measures and dimensions. The typical example is a cube to analyze sales, where the analyzed fact is usually the commercial transaction (what was sold, when, at what price, etc.):

  • Dimensions are the attributes that let users sort out the fact being analyzed. In the case of sales, they could be the customer, the date, the product, the salesperson, etc.
  • Measures are the attributes that quantify the fact being analyzed. In the sales example they could be the number of units sold, the price, the total invoicing, etc.

Dimensions

Dimensions establish different forms of sorting out the facts being analyzed. The most typical dimension is the "Date", which sets the fact in time. Other typical dimensions are Products, Customers, Locations, Salespeople, etc.

Dimensions are organized in hierarchies (i.e. Year / Quarter / Month) to make drilling the data easier during analysis, given that the most frequent queries are those geared towards showing more detail (drill down), for instance, breaking up a year in its quarters, or quarters in their months.

Measures

Facts considered for each model are those directly involved in the topic the model focuses on. Measures (sometimes only known as facts) are values that quantify an aspect of the reality being analyzed.

Measures quantify independent facts, but are normally added to quantify a set of related facts. The typical aggregation function is the addition but there are a number of possibilities (averages, minimums, maximums, etc.)

Hierarchies in the Dimensions establish the natural order of aggregation. Thus, for instance, monthly sales may be analyzed in an aggregated way for a whole quarter or a whole year.

Elements of the Dimensions selected for a Query

O3 Browser information analysis is based on direct queries made by the user and that are mainly determined by which elements being consulted in each dimension are.
The user has a control panel in the Dimensions Bar that states which elements you wish to include in the query for each dimension in the model. This graphic control enables different ways to determine which the elements to consult are.

The following are the different type of queries the user can explore:

  • Explore Children:
    "Children" elements to a given element are shown. For example, in a date type dimension, if we select the first quarter of 2005, the "Explore Children" query shows the 3 months of the first quarter on the corresponding axis.
  • Single Element:
    Only the selected element is shown, regardless of the fact it may have lower levels in the hierarchy of the dimension.
  • List of Elements:
    The list of selected elements is shown, which must all belong to the same level of hierarchy. For instance, in a date type dimension you may select all the first quarters every year.
  • Drill level:
    All elements in the hierarchical level selected are shown. For instance, in a date type dimension, the query "Drill quarter level" shows the quarters corresponding to each of the years for which there are data.
  • Advanced operations including the following options:
    • Level
    • Descendants
    • Siblings
    • Leaves
    • Ancestors
    • Father
Unknown macro: {scrollbar}