I studied architecture in middle school (yes, I am that big of a nerd) and fell in love with the patterns of Frank Lloyd Wright designs. Widely regarded as one of the best architects in U.S. history, Wright’s designs are legendary for their attention to the smallest of details. He drew the plans, selected the materials, designed the window panes, built home specific furniture, etc.
For example, “As he did with all his residential architecture, Wright gave the Pope-Leighey House a horizontal orientation. This effect is established by the flat rooflines, and the idea extends to the way the bricks are laid. The mortar in the vertical joints is tinted to match the bricks, so that they recede, enabling the white mortar in the horizontal channels to visually pop out. This allegiance to a horizontal orientation is so all-encompassing that even the screw heads (every one of them) that attach the cypress to the plywood are perfectly aligned horizontally.”
Data is no different. Just as Wright thought about each and every detail and made decisions on how it related to the objective of the whole design, so should a data architect. After all, each row, column, and table added will impact the overall design’s ease of use and its performance. To continue the analogy, the data model represents the drawing of your house. The technologies used for input, storage, access, and extraction should all be considered as well. Oracle and Hadoop are different, as are SAS and Qlik. The Database technology chosen is akin to the materials used to build the house. Whatever the need, you want to design it optimally for the technologies you plan to use. After all, a high-rise in New York city isn’t built out of brick – for good reason. The front end BI tools are the equivalent of the interior design. It’s their job to tie it all together, make it pretty, and allow for it to be used.
Anyone can draw and build a house (or a data model), but it takes an architect to make it art.