HughCraig / GHAP

1 stars 0 forks source link

Consider grid and region representations of data #304

Open BillPascoe opened 10 months ago

BillPascoe commented 10 months ago

In a different post Hugh had suggested using DGGS. We should consider how this or other grids could be applied to Australia. We should also consider whether to assign LGA, SA2, or other kinds of region to the data, or represent it that way, or allow for export that way, etc?

BillPascoe commented 10 months ago

Hugh's previous post on a different issue: DGGS may be useful for representing areas. According to [https://www.ogc.org/projects/groups/dggsdwg]:

Such a common framework exists via Discrete Global Grid Systems (DGGS), although DGGS are not yet widely understood or adopted technologies throughout the geospatial community. A DGGS is a form of Earth reference that, unlike conventional coordinate reference systems (which represent the Earth as a continual lattice of points), represent the Earth using a hierarchical tessellation of equal area cells; where each cell is referenced by a globally unique identifier. A key differentiation of DGGS from other GIS technologies is that a DGGS represents the entire curved surface of the Earth rather than a flattened map projected representation. This means that spatial analytics conducted within a DGGS do not need to account for the added distortions created when map projections are applied to data; thus, leading to more efficient and more accurate outcomes. Because each cell of a DGGS represents a uniform spatial unit with a unique reference identifier, a DGGS is able to reduce complex multi-dimensional spatial queries into one-dimensional array processes that enable rapid and accurate computation and integration of data.

While conventional coordinate reference systems are designed to facilitate repeatable navigation, a DGGS is designed to ensure a repeatable representation of measurements – observations, interpretations, and events. Every item of information in a DGGS is associated with an area, and spatial resolution is explicit. This is much preferable to tagging an attribute with a latitude and longitude, since it's never clear what area possesses the attribute, or how accurate the measurement of location is. Combining or integrating layers becomes trivial in a DGGS, because items of information are automatically spatially associated with each other. This is much like overlaying information across congruent rasters, and far easier than having to perform overlay using points, lines, and arbitrary areas.

HughCraig commented 7 months ago

OK, pretty speculative at this point.