Data Modeling, An Overview

Data Modeling Concepts

The word data originates in mid-century Latin, stemming from the word datum and originally meaning “something given” or granted [1]. Philosophically, we might think of this as whenever we obtain a set of information, we have been given the opportunity to gain knowledge. The first time the word data appears in English, it immediately evokes the need for a data model; ‘From all this heap of data it would not follow that it was necessary…” [2]. This is in line with the Oxford English Dictionary’s definition, “a thing given or granted; something known or assumed as fact, and made the basis of reasoning or calculation; an assumption or premise from which inferences are drawn” [2].

A data model is a specified definition of a framework for structuring data and formatting it into an organizational information system. A data model generally comprises instructions for data connections, inputs, outputs, data flows, loops, processing and storage. Data models illustrate the structure of data, organizing it such that it can support the analytical needs of decision makers. Well-constructed data models are able to be replicated because at its core, a data model is a set of instructions.

During the data modeling process, analysts envision plans for unorganized data by simplifying data into their smallest units. These small data units are then charted by their relationships to the other small data units to recreate a unified big picture. The analysts’ goal is to make the data pipeline frictionless for end users. A good data model result is simple, flexible and organized for consistent human interpretation, so that when followed the model should produce consistent and reliable results.


Misconceptions

  • A data model is frequently mentioned as the first steps in building out a database, leading many to think that is its only purpose. This is not the case. A data model may define a database; however, it is not a database, nor limited to being used solely for the construction of databases.
  • Data models describe the structure of data in a system, but they are not limited to structured data. Unstructured data, although far less typically, may also be described using data models.
  • A data model may be illustrated in a graphical form, such as a flowchart, but it is itself, not a flowchart.

Conceptual Diagram (Concept, Not Technology)

Conceptual diagrams and schemas typically target high-level descriptions of business rules, concepts or experiences. In the example [Image 01], is a conceptual diagram for a blended reality distributed system. It describes the management of synchronized multiple dual reality states. From this diagram, we can see that a change in any virtual object of a given Interreality Portal results in identical changes to all connected Interreality portals [3]. Further, a change in any xReality object of a given Interreality Portal also results in changes in the representation of the real device on all subscribing Interreality portals [3].


Logical Diagram (Data Flow/Structure, Not Hardware)

A logical data diagram and schema tends to focus on a specific problem domain.  Typically, logical diagrams describe data structures, for example; relational tables and columns, object-oriented classes, and/or XML tags, however, still at the higher levels similar to conceptual models. Logical diagrams follow the business requirements. In this example [Image 02], we see a data flow for an example application in the area of Virtual and Augmented Reality based on VRML/X3D [4]. We can see Video coming in and going to the ARToolKit where frames and projections are piped into the VRML Scene and M1, M2 and related components are piped into the MarkerGoup, then into the VRML Scene once they have been grouped.


Physical Schema Diagram

Physical schema diagrams are designs for actual information systems, complete with primary keys, foreign keys and constraints. These models need to have extra care taken while being creative, since naming conventions and such require a high degree of accuracy. Physical models are typically transformed from logical models. For this example [Image 03], we are looking at the dimensional model the data team constructed for the upcoming post production workflow project. This diagram lays out the details of one of the databases.


Entity Relationship Diagram

An entity relationship model in its simplest form is essentially an abstract conceptual/semantic model, which defines structed data. In this example [Image 04], we see a Chen’s notation entity relationship diagram of a Massively Multipplayer Online Role-playing Game (MMORG). We can see the entities; account, region, character, item, item instantiation, creep instantiation, creep, etc… connected to mostly binary relationships. For instance, is the IDNum, a creep 1 (=yes)(N=no)… it it carrying… is it containing any connections to regional attributes, such as the regional name etc…


Data Modeling and Analysis for the AR/VR Industry

In 1997, the first recognized use of the term “big data” appeared in a paper by NASA scientists describing “a problem they had with visualization (i.e. computer graphics)” [5]. They noted that the “data sets [were] generally quite large, taxing the capacities of main memory, local disk, and even remote disk” [5]. Sound familiar?  My favorite part of the paper is the attitude of the scientists, which is truly the take away. They saw that big data provided “an interesting challenge for computer systems” and then began trying to address it. Again, sound familiar…?

The VR industry is riddled with big data problems that data modeling and analysis can help to solve. Like the NASA scientist, mentioned earlier addressing these issues is a matter of attitude. We are committed to working together to implement new solutions and make sure everyone’s voice is heard throughout the data modeling project. These last examples are a few of the diagrams we see and use every day [Images 567 ]. Hope this post was informative and that you all have a wonderful week full of adventure.


References

[3]. Callaghan, V. (2013). XReality Interactions Within an Immersive Blended Reality Learning Space. Retrieved May 31, 2017, from http://www.academia.edu/12416218/xReality_Interactions_Within_an_Immersive_Blended_Reality_Learning_Space

[2]. Gray, N. (n.d.). Data is a singular noun (S. Draper & P. Coles, Eds.). Retrieved May 31, 2017, from http://nxg.me.uk/note/2005/singular-data/ Blog of Norman Gray, Researcher, Astronomy Group, School of Physics and Astronomy University of Glasgow, Glasgow, G12 8QQ, UK

[1]. Harper, D. (n.d.). Data (n.) (The Online Etymology Dictionary, Ed.). Retrieved May 31, 2017, from http://www.etymonline.com/index.php?term=data

Press, G. (2014, November 17). 12 Big Data Definitions: What’s Yours? Retrieved May 31, 2017, from https://www.forbes.com/sites/gilpress/2014/09/03/12-big-data-definitions-whats-yours/#1778924313ae

[4]. Seibert, H., & Dähne, P. (1970, January 01). JVRB – Journal of Virtual Reality and Broadcasting. Retrieved May 31, 2017, from https://www.jvrb.org/past-issues/3.2006/777

Dawn

Leave a Reply

Your email address will not be published. Required fields are marked *