Four principles of data intelligence

Indicators

A decisional indicator is a numerical variable the can be collected in a running operational information system and can be calulated. The decisional system is a second independant information system that observes an operational information system in order to produce sufficiant information for be able to make management decisions over the activity. Tis observation is mostly done using numerical and computable values (indicators) fromwhich statistical summations can be calculated on several ranges and scales of observation. Computing those indicators to output valuble statistics with functions such as SUM, MEAN, MAX or MIN is called an aggregation.

Dimensions and axis

Indicators are NUMBERS colected from the information stored within the information systems that is being observed (Moodle for the TOA for Moodle product). The collected values are uually obtained within a certain context. They are originated fromsome precise point of the system and are linked to other information that qualify the extracted value. Binding a numerical measurement to other information wlll allow creating a context in which this collected information can be used. E.g., if some measurement of a use time is bound to a unique user, this user will be the surrounding context of validity for this information, and will qualify this observed duration as "the time THIS user spent". Here shold we introduct the definition of a dimension of analysis, allowing separate counting of the times spent by each user, or conversely summing the time for the whole "users" dimension.

Going further, a user to which this time was attributed could be himself bound to a certain city through his profile information. The base information (a use duration measurement) is also indirectly attached to the city of the user. "user" and "city" dimensions are attached by a dependancy.

A serie of successive dependancies form a chaîn and assembles dimensions within an analysis axis. The Trimane Open Analytics pour Moodle tool only yet knows about linear axis that have no mulitple derivations in it.

In an axis, the dimension that is closer to the unity measurement is called the base dimension.

Data Cube (hypercube)

To be translated

Report

Back to TOA summary

Last modified: Sunday, 14 February 2010, 06:11 PM