Geographic Information System - Assignment Example

www.allwriting.net Sumanta Sanyal Dated: 04/09/07 Geographic Information Systems Introduction: This paper contrives to answer eight questions specifically set on particular aspects of the ‘Geographic Information Systems’ (GIS). The paper now proceeds forthright to the questions and their answers. “GIS is an information system”: Question 1: GIS can be seen or viewed from different perspectives. Discuss the view that ‘GIS is an information system’. Give an example that supports this view. Answer 1: The ESRI (Environmental Systems Research Institute, Inc.) describes the GIS as – “a collection of computer hardware, software, and geographic data for capturing, managing, analysing, and displaying all forms of geographically referenced information” (What is GIS?, ESRI, 2007). It becomes clear from this description that GIS is not a specific technology but a set of technologies that primarily yields geographically referenced information for various needs. Thus, it is clear also that three broad sets of technology are involved – 1. computer hardware, 2. computer software and 3. geographic data producing technologies. Using an example already posted on the ESRI website it becomes more evident that GIS indeed is an information system. In this example it is posited that Bank of America wanted to configure its branch coverage in relation to the deposit potential in the New York area. The paper expands on the example to elucidate better. To analyse its true coverage the bank would require to first posit its existing branches in the area geographically, probably on a map. Next, it would need to analyse which parts of the map display the most number of branches and which parts the least. In this manner the bank will be able to determine how exactly it is positioned in relation to the deposit potential in the area. Thus, areas where it has relatively large numbers of branches would be well covered while areas where the branch numbers are relatively low would be weakly covered. Expanding further on the example, it can be imagined that GIS can be used to allow the bank to configure where it should place (the areas with relatively affluent populations) the most number of branches and where it should use caution (the areas with relatively low income populations) in positioning branches. GIS helps the bank in the above example by providing the required information – either available internally within the bank or externally at some such information providing agency. This is a typical GIS task and it clearly enough delineates that GIS indeed is an information system and not a technology or any other form of institution. It is an information system that utilises a set of technologies to yield geographically referenced information. Topology – A Data Structure Concept in GIS: Question 2: Topology is one of the most useful data structure concepts in GIS. Define and explain (or use illustrations, if needed) the term topology. Discuss its importance in GIS. Answer 2: The concept of ‘topology’ was mathematically developed by Leonhard Euler in 1736 (Theobald, 2001). Mathematical topology describes geographic planar space as being two-dimensional – latitude and longitude. Since human populations reside on the Earth’s planar surface this is germane. GIS defines ‘topology’ as – “the spatial relationships between adjacent or neighbouring features” (Theobald, 2001). This definition allows the planar data to be distributed into areas with defining lines that intersect at nodes. The areas are polygons (two-dimensional cells), the lines are edges or arcs (one-dimensional cells) and there are the nodes (one-dimensional cells). This is as per figure 1 below (Theobald, 2001). Fig. 1 (Source: Theobald, 2001) A typical GIS generated data structure utilising topological features is the ‘ArcInfo’ structure where topological relationships between adjacent polygons can be ‘explicitly’ stored in the Arc Attribute Table (AAT) by storing the adjacent polygon IDs in the LPoly and RPoly Fields (L-left; R-Right) (Theobald, 2001). Adjacent lines connect at nodes and this information is stored in the arc node Table (Theobald, 2001). There are a number of advantages to this manner of topologically generated data structures. These are as below. It automatically facilitates digitalising and management of editing errors and artefacts. It reduces data storage for different polygons as the adjacent boundaries are stored only once. It facilitates advanced spatial analyses like adjacency, connectivity and containment. Since the topological structure has space-filling polygons that do not overlap, other usual cartographic features are not essential to these forms of data structures reducing storage space. (Theobald, 2001). For all the above reasons mainstream GIS software do not use cartographic (non-topological) data structures (Theobald, 2001). The ‘Raster’ Versus ‘Vector’ View: Question 3: Discuss the main conceptual differences between the ‘raster’ and ‘vector’ views of spatial objects. List and discuss three (3) advantages of raster data structure over vector data. Which data structure will you use in producing road/street map products for drivers? Why? Answer 3: The GIS system represents the real world with either one of two spatial models – raster and vector (Husdal, 1999). It is fortunate that this paper has already treated in part these two models in the previous section on topology. The arc-node table used to represent topology uses nodes, arcs and polygons – incorporated into zero-, one-, and two-dimensional cells respectively. This is the exact raster model of spatial representation (Husdal, 1999). Conversely, the represented spatial objects – intersections by nodes, lines by arcs and the delineated areas by polygons – the intersections, lines and delineated areas, when converted from the raster version become the spatial vector representation (Husdal, 1999). Thus, the raster data, when graphically represented, is illustrated by figure 2. Conversely, the vector data, when graphically represented, is illustrated by figure 3. Fig. 2 Fig. 3 (Source: Amherst College, 2007). While the above representations are rather simplistic the general impression can be derived that the raster model represents real-world objects as grid systems of rows and columns of cells while the vector model represents objects as they are in the real world when representation is geometric. The raster model illustrates other related data such as terrain topography and demographic characteristics by varying cell colours and resolutions (Amherst College, 2007). The vector model does so by using a variation of points, lines and polygons (Amherst College, 2007). Three advantages of the raster model are as follows: the data structure is simple; the model is ideal for remotely-accessed or scanned data; and procedures for spatial analyses are simple. (USGS, 2006) It is notable that is road/street map products for drivers have to be prepared the vector model may yield aesthetically pleasing and life-like graphics but the high costs of the hard- and software and the complex data structures would pre-empt its usage in a small organisation, presumably me. Thus, I would prefer to use the raster model for preparing the map products because I think the drivers would compromise on the aesthetic quality in exchange of lower costs of the products (USGS, 2006). Map Projections in Spatial Data Handling: Question 4: Why do we need to understand the concepts of map projections in spatial data handling? Why is it best to know more (rather than less) about it, if you are working for an international consulting company that handles various projects around the world, and may need to integrate various datasets from different countries? Answer 4: In GIS spatial data representing areas on the Earth have to represent the Earth or parts of it on a flat surface like a computer screen. This flat representation of the Earth is generally called a map and the flattening process is called projection (Finding Data for Maps, Amherst College, 2007). There are three basic types of flattening processes – planar, conic and cylindrical (Finding Data for Maps, Amherst College, 2007). Unfortunately, all projections tend to distort some aspect of the Earth’s geometry. They do so in the following manners: all projections distort distance to some extent; some projections are capable of conserving angles and, thus, small shapes; such projections are said to be conformal; and some projections can conserve relative area and are said to be equal-area. (Finding Data for Maps, Amherst College, 2007). Unfortunately again, no projection is capable of being both conformal and equal-area. Thus, famously, the Mercator map projection conserves angles and is said to be conformal but it severely distorts area, especially near the poles. If I am working for an international company that needs to integrate its datasets from around the world I would choose conformal projections for a single large and global dataset while I would choose equal-area projections for more localised and smaller datasets. This is because it would be difficult to project the entire globe or even large parts of it through equal-area projections that, on the other hand, would benefit small area maps with greater conformity to actual local areas. Digitising Vs. Scanning: Question 5: Differentiate digitising from scanning. Compared with manual digitising, scanning is generally considered as quick and less costly. Identify at least two situations or conditions where manual digitising could be more preferable than scanning. (You may want to focus on the characteristics of the data to be captured.) Answer 5: When planning to construct a GIS database system it is necessary to understand that critical layers that need to be generated have to be identified and the levels of accuracy of eaxh layer has to be determined to enable successful and efficient operation and analyses (Ramsey, Undated). Digitising is a process that entails manual input of data in the forms of the likes of orthophotos and transparencies that have to be further modified and enhanced manually at the computer to promote them into spatially oriented GIS datasets (Ramsey, Undated). Such a process may be time-consuming but, effectively, there is a debate going on the cost structures of digitising where manual handling is opposed to automation (Ramsey, Undated). It is notable that digitising is primarily vector model datasets. In contrast, scanning is efficient and easy as well as time conserving yet its raster modelled structure, as with other raster models, leaves much to be desired in the dataset features where such features are deemed necessary (Ramsey, Undated). Thus, while manually digitising may seem time-consuming and, often, primitive, it is still the preferred method in cases where vector models are required for constructing certain datasets within the overall GIS system. The scanning method is preferred where no special need arises for the special features available only in the vector models (Ramsey, Undated). The Scale of Measurement: Question 6: For an attribute data, the level of measurement (also called 'scale of measurement') could be nominal, ordinal, interval, or ratio. Give three (3) examples of attributes (e.g. soil type, elevation, etc-) that could be represented by each scale of measurement. (Three examples for each category). Explain the importance of the scale of measurement in handling and analysing spatial data. Answer 6: Data in the real world is represented in GIS databases as either continuous datasets or discrete ones (Ramsey, Undated). Continuous datasets may represent that reality which is expected to vary spatially, temporally or thematically (Ramsey, Undated). Geographic features like temperature, elevation and barometric pressure vary across the Earth from place to place (spatially) and time to time (temporally). Thematic variations may be in demographic characteristics of populations across the Earth – educational levels, socio-economic status, etc. As per vector models, such variations may be depicted by intensity of points and other illustrative media while, in the raster model, such variations are depicted by varying the cell pixel values (Ramsey, Undated). The pixel variations are at the best approximate and reality cannot yet be fully depicted by this graphic rendition. Discrete items are fixed and characteristic of a particular Earth locality and capable of being depicted in the vector model points, lines and polygons though the raster model can also be used (Ramsey, Undated). Features like roads and buildings can be depicted in this manner but require a universal scale of measurement to preserve uniformity across various datasets when they must be put to use simultaneously (Ramsey, Undated). Thus, GIS datasets are constructed on layers of data that may have four possible scales of measurement – 1. nominal – names, labels or numbers without any special mathematical relationship among them; 2. ordinals – attributes of features can be built up in a sequence where there is no definite mathematical relationship but it can said that one is greater than the other; 3. interval – here features are placed at regular intervals where the intervals have mathematical significance but the sequence does not start at zero – example – temperatures in the Celsius and Fahrenheit scales; and 4. ratio – here the features are also placed at regular intervals but the sequence starts at zero – example – temperatures in the absolute scale. (Ramsey, Undated) Discrete items require nominal and ordinal measurements – example: buildings, roads, rivers, etc. – while continuous items require interval and ratio scales – example – temperature, pressure and elevation (Ramsey, Undated). It is obvious now that scaling data makes variant items within it distinct and facilitates handling and analysis. GPS and Remote-Sensed Data: Question 7: Differentiate (a) GPS derived data from (b) remote sensing data in terms of the technology involved and the nature of the data they produce for GIS use. Which one produces data in raster structure? Explain. Remote sensed data is acquired through sensors (cameras or any other devices) that can record electromagnetic radiations (EMR) emanating or reflecting from spatially removed objects. Such sensor devices record EMR by variations in the radiation characteristics (frequency or wavelength etc.) (Raber et al, 2005). Such recording is done either in the analog or digital modes. In the analog mode the variations are themselves recorded on some storage device like photographic film while in the digital mode these variations are ascribed numerically in the binary system (Raber et al, 2005). It is notable that data acquired through remote sensing tend to be distorted. Since, for GIS, most data is acquired from high above ground the distortions tend to increase away from the lowest point – the nadir (Raber et al, 2005). Also, the data is not spatially oriented and geographic identity is not available. To correct this GPS system is used to earmark the object area with control points that are easily identifiable geographically and, when recorded separately, can be mapped onto the remote sensed data to create geographically sensible (georectified) data that can be inducted easily into the GIS database (Raber et al, 2005). It is thus noted here that while the GPS system records data in the vector format the remote sensed devices record data more specifically in the raster format. Mapping the GPS data onto the remote-sensed one georectifies it and makes it available to GIS databases. It is also notable that data acquired in the remote-sensed manner is essentially a record of EMR from desired objects and this is a rather simplified version of the technology. Answer 7: Database Management Systems (DBMS) in GIS: Question 8: As a technology, database management systems (DBMS) evolved separately from GIS. Discuss the importance of DBMS in a GIS. Why 'common' or typical DBMS software like Microsoft Access could not replace the data management functions offered by a GIS? Answer 8: DataBase Management System software did not originally incorporate data with spatial configurations. Nevertheless, when it became apparent that GIS is a particular favourite among users, especially those in need of data with such spatial configurations, DBMS systems began adapting to the new requirements (Bruenig and Zlatanova, 2004). The new GIS users wanted multi-user access to data stored in DBMS systems and this meant that the DBMS system had to adapt and the adaptation was wrought in two ways. Top-down approach – The GIS users started to store non-spatial data like thematic content that they could access anytime in the system. This is the top-down approach since the DBMS system is constructed under the guidance of the GIS one (Bruenig and Zlatanova, 2004). Thus, the GIS system can selectively access the geodata stored in the DBMS system. Bottom-up approach – In this approach the DBMS system is constructed in such a manner that full sets of spatial data is available within in additional to what may be non-spatial one. The GIS system can now readily access whatever it requires within the larger DBMS system (Bruenig and Zlatanova, 2004). Nowadays, most DBMS systems store full type geodata for usage by GIS systems. Since data management systems like Microsoft Access are geared to handle non-spatial data they cannot replace the GIS system that specifically deals with spatially oriented datasets (Bruenig and Zlatanova, 2004). Also, since this new development where multiple users access data stored in DBMS systems, the DBMS system has been upgraded to one that can implement multi-user control and crash recovery to pre-empt spatial data corruption due to multi-user handling (Bruenig and Zlatanova, 2004). DBMS systems already had such protection features for non-spatial data and it is good that GIS systems can avail of this additional facility when associated with the DBMS databases (Bruenig and Zlatanova, 2004). References: What is GIS? Environmental Systems Research Institute, Inc., 2007. Extracted on 1st September, 2007, from: http://www.gis.com/ Introduction to GIS, Amherst College Information Technology, 2007. Extracted on 1st September, 2007, from: http://www.amherst.edu/it/software/gis/gis-introduction/#ABriefOverview Finding Data for Maps, Amherst College Information Technology, 2007. Extracted on 1st September, 2007, from: http://www.amherst.edu/it/software/gis/gis-introduction/#ABriefOverview GIS Data Structures, Northern Prairie Wildlife Research Center, USGS. Extracted on 2nd September, 2007, from: http://www.npwrc.usgs.gov/resource/habitat/research/struct.htm Husdal, Jan, Network analysis – network versus vector – A comparison study, 1999. Extracted on 2nd September, 2007, from: http://husdal.typepad.com/blog/1999/10/network-analysi.html Theobald, David M., Understanding Topology and Shapefiles, 2001. Extracted on 1st September, 2007, from: http://www.esri.com/news/arcuser/0401/topo.html Ramsey, R. Douglas, Representing Reality in a GIS and Data Input, Undated. Extracted on 2nd September, 2007, from: http://www.nr.usu.edu/Geography-Department/rsgis/GIS/Lectures/datinp/datainp.html Raber, George, et al, Remote Sensing Data Acquisition and Initial Processing, Earth Observation Magazine, 2005 July, Vol. 14, No. 5. Bruenig, Martin, and Zlatanova, Sisi, 3D Geo-DBMS, Directions Magazine, 2004. Extracted on 2nd September, 2007, from: http://www.directionsmag.com/printer.php?article_id=694 Read More