Hay_Topic 11 - Vector analysis_v1 (1).pdf

Geog 380: Geospatial Communication Source: https://earth.nullschool.net Topic 11: Spatial Analysis - Vectors GEOG 380 - Topic 11 © Geoffrey Hay (2023) 1 Learning outcomes By the end of this lecture topic and associated labs, a successful student will be able to: § Recognize some of the option...

Geog 380: Geospatial Communication Source: https://earth.nullschool.net Topic 11: Spatial Analysis - Vectors GEOG 380 - Topic 11 © Geoffrey Hay (2023) 1 Learning outcomes By the end of this lecture topic and associated labs, a successful student will be able to: § Recognize some of the options and strategies for three major categories of GIS analysis operations: (i) queries, (ii) measurement, and (iii) transformations § Perform some simple overlay and buffering operations GEOG 380 - Topic 11 2 Processing in Raster & Vector § New geospatial technicians often think of raster and vector data analysis as two different worlds § In reality, most software packages provide relatively seamless integration between the two (Spatial Analyst in ArcGIS Pro) § Which is better? http://gsp.humboldt.edu/olm/Lessons/GIS/08%20Rasters/ RasterToVector.html GEOG 380 - Topic 11 § Actually, most operations can be accomplished with both data models § However, conversion between the two can introduce errors § Often best to stick with the “native” data format 3 Problems with Conversion § Whenever you convert data from raster to vector or vector to raster, you will degrade the data by some amount. Issues include: https://vimeo.com/53016715 § § § § GEOG 380 - Topic 11 Loss of detail Loss of accuracy Stair stepping (raster to vector) Changes to the original data 4 Choosing Between Data Models “Raster is faster but vector is corrector” § Raster data is useful when: § Working with continuous data types: elevation, slope, satellite photos § Good for large area analyses § Good for surface analysis § Mathematical modeling § Spatial detail isn't important GEOG 380 - Topic 11 § Vector data is useful when: § Working with discrete data types: trees, buildings, property boundaries § Good for small study areas § Spatial detail is important (When "close enough" isn't really good enough) § When topology is needed for the analysis 5 Choosing Between Data Models § Other factors that can influence your decision include: § Available storage: some rasters are really large (think Petabytes = 1000 TB) § “Some estimates hold that a Petabyte is the equivalent of 20 million tall filing cabinets or 500 billion pages of standard printed text” - Google § Expected types of analysis: some tools only work with raster or vector data § Expertise of human operators § Level of accuracy desired GEOG 380 - Topic 11 6 Conversions § Raster to Points: § Raster to Point § Raster to Polyline: https://www.onestopgis.com/GIS-Theory-and-Techniques/Raster-DataAnalysis/Comparison-of-Vector-and-Raster-based-Data-Analysis/1Comparison-of-Vector-and-Raster-Based-Data-Analysis.html § § § § Contour Stream to Feature Raster to polyline Cost Path § Raster to Polygon § Viewsheds § Watersheds § Raster to Polygon GEOG 380 - Topic 11 § Point to Raster § Interpolation tools: Kriging, IDW, Splines, etc. § Density tools: Kernel density § Point to Raster § Polyline to Raster § Polyline to Raster § Polygon to Raster § Polygon to Raster § Topo to Raster 7 GIS Analysis Terminology § Entity: An individual point, line, or polygon in a GIS database § Attribute: Data about the entity § In a vector GIS, attributes are stored in a database, and there may be many attributes for any given entity § In a raster GIS, the grid code (x, y) represents the spatial attribute for the pixel, and there typically is only one (non-spatial) attribute per pixel i.e., height, brightness, temperature, etc. GEOG 380 - Topic 11 8 … continued § Data layer/coverage: one of a series of themes or data sets in a GIS, consisting of entities and their associated attributes § Note: the general term ‘coverage’, as used here, is actually a generalization of the original term, which refers to the Arc/Info native file format. However, it’s used so often in this context (data layer) that I’m including it. https://vimeo.com/53016715 § Image: a raster data layer § Note that images (rasters) are not limited to remote sensing imagery. Anything that we store ‘pixel by pixel’ is a raster GEOG 380 - Topic 11 9 GIS Analysis Terminology, Cont… § Function/operation: a data analysis procedure performed by a GIS § Algorithm: an unambiguous sequence of functions/operations designed to solve a problem GEOG 380 - Topic 11 10 Queries § Queries: questions posed to the database that do not involve changes to the database. § There are many ways of interrogating the database for information in formats that would qualify as queries § § § § § ‘Right-click: properties’ on individual layers Cursor coordinates ‘Identify’ mode Histograms, scatterplots, and other charts/graphs Formal database queries GEOG 380 - Topic 11 11 … continued § Two basic types: spatial and aspatial § Spatial queries operate “by location” (found under the Selection menu in ArcGIS Pro) § Select features that intersect… § “ ” that are within a distance of… § “ ” that are completely contained by… § “ ” that contain… § Etc… GEOG 380 - Topic 11 12 … continued § Aspatial queries interrogate the attribute tables through structured query language (SQL) § SELECT … FROM … WHERE … § Found under the Selection menu in ArcGIS Pro GEOG 380 - Topic 11 13 Boolean Expressions § § Boolean logic is based on the evaluation of binary values (yes/no, on/off, 1/0) Complex queries can be created by combining individual expressions (spatial or aspatial) through boolean operators: § § § § GEOG 380 - Topic 11 AND: true if both are true OR: true if either is true XOR: true if exactly one is true NOT: inversion expression 14 Measurements § Measurements: simple measurements regarding geographic entities or attributes § The ability of computers to make accurate and consistent area measurements was the original motivations for the Canada Geographic Information System (CGIS): the world’s first GIS https://www.esri.com/arcgis-blog/products/arcgis-earth/3dgis/perform-interactive-measurements-in-arcgis-earth/ § Common GIS metrics: length, perimeter, and area § Also, many others (shape, aggregation, interspersion, etc.), though not as common GEOG 380 - Topic 11 15 Distance Measurements § Simplest method of determining distance between points in space is the Pythagorean or Euclidean distance § Distance between A and B defined by hypotenuse of a right-angle triangle § Assumption is that the points lie on a flat plane, and, of course, the world is not flat! A B GEOG 380 - Topic 11 16 … continued § Great circle distance calculates the metric for a spherical earth § Again, though, the earth is not completely spherical (geoid!), so even more complex procedures may be called for under certain circumstances https://users.cs.jmu.edu/bernstdh/web/common/lectures/su mmary_great-circle-distance_spherical.php GEOG 380 - Topic 11 17 … continued § Manhattan distance calculates the distance along raster cell sides § “The distance between two points measured along axes at right angles.” § This could result in a length distortion created by the raster data model vs a vector model GEOG 380 - Topic 11 18 Distance Measurements, Cont… § Remember, ALL measurements in GIS are an approximation, due to several factors ESRI online help § Vector: lines and polygons are made up of straightline segments (polylines) § Raster: all entities (points, lines, polygons) are represented with pixels (i.e., small squares) § Distortions inherent in the projection, and because of slope, and elevation differences. § Generalization of complex objects by simple models GEOG 380 - Topic 11 19 Area & Perimeter Measurements § GIS perimeter estimates are a fairly straightforward extension of distance measurements § Summing up the lengths of the polylines defining the polygon in a vector layer § Multiplying the number of cell sides that make up the boundary of the feature by the pixel size § Similarly, area measurements are relatively simple extensions of geometric logic https://rdkb.sgrc.selkirk.ca/Help/Content/Client_APIs/SV_User/ SVU_MeasureDistanceOnTheMap.htm GEOG 380 - Topic 11 20 Transformations § Transformations: simple methods of spatial analysis that transform attributes or entities of geographic data into useful NEW products https://desktop.arcgis.com/en/arcmap/latest/manage-data/using-sqlwith-gdbs/spatial-operation-functions-for-st-geometry.htm GEOG 380 - Topic 11 § Buffers, overlays, point-in-polygon, interpolation: these relatively simple operations form the basis of powerful algorithms that solve a multitude of geographic problems 21 Transformations: Buffering https://saylordotorg.github.io/text_essentials-of-geographic-information-systems/s11-01-single-layer-analysis.html § A buffer operation builds NEW objects around existing point, line, or polygons by identifying all areas within a specified distance § Among the most useful and popular GIS operations § Identifying danger zones in an industrial accident § Estimating customers within driving distance of a retail store § Avoiding protected riparian areas for logging operations GEOG 380 - Topic 11 22 … continued § Buffering can be done with both raster and vector layers, but rasters enable the simple extension of buffering to consider factors other than simple distance § E.g., buffering access to a city in a manner that incorporates travel speed, by assigning a friction cost to each cell http://www.geog.ucsb.edu/~good/176b.old/10.html#1 GEOG 380 - Topic 11 23 Transformations: Point in Polygon § The point-in-polygon operation determines whether a point lies inside or outside a polygon, though it can be extended to consider many points and many polygons § Points could represent instances of disease in a population; polygons health districts. Which districts are affected by a particular outbreak? § Points could represent rare animal sighting locations; polygons represent management districts. What areas should be subject to special species-at-risk regulations? GEOG 380 - Topic 11 24 Transformations: Point in Polygon § Standard algorithm for determining whether a point lies within a polygon involves extending a line from the point towards infinity § If the line crosses the polygon’s boundary/edge an odd number of times, or if it lies on the polygon edge: it is in § If it crosses an even number of times (or zero): it’s out http://www.geog.ucsb.edu/~good/176b.old/10.html#1 GEOG 380 - Topic 11 25 Transformations: Polygon Overlay § Polygon overlays involve the spatial intersection of two different polygon layers, and are among the most varied and powerful families of GIS operations http://www.geog.ucsb.edu/~good/176b.old/10.html#1. § What proportion of a protected area was affected by a forest fire? § What is the total area of forest owned by each property holder? GEOG 380 - Topic 11 26 ESRI online help Overlay operations GEOG 380 - Topic 11 27 Polygon Overlay Issues § § Polygon overlay operations with vector datasets are problematic because of the potential for slivers being generated by the procedure The same boundary (e.g., a coastline) may be represented slightly different in each layer, because of slight digitizing errors § Paradoxically, the more detail you put into the digitization, the more pronounced this problem becomes! https://www.onestopgis.com/GIS-Theory-and-Techniques/Spatial-Data-Editing/TopologicalErrors/2-Types-of-Topological-Errors-with-Geometric-Features.html § Most software packages try to minimize this problem though tolerance thresholds, but these must be set correctly Ø Not an issue in raster overlays, since the entities (pixels) are all the same GEOG 380 - Topic 11 28 Transformations: Spatial Interpolation § Spatial interpolation: algorithms that attempt to estimate values of a continuous field in places where no measurements exist in the database § Estimating rainfall in places away from individual weather stations § Estimating elevation from contour lines https://gisgeography.com/kriging-interpolation-prediction/ GEOG 380 - Topic 11 29 https://www.angelchen.net/blog/2018/1/11/spatially-interpolating-snow-water-equivalent-data-in-the-canadian-prairies Transformations: Spatial Interpolation § Spatial interpolation: algorithms that attempt to estimate values of a continuous field in places where no measurements exist in the database § Estimating rainfall in places away from individual weather stations § Estimating elevation from contour lines § GEOG 380 - Topic 11 Thiessen polygons, inverse-distance weighting, kriging: topics for another day! 30 GIS Analysis: Summary § Queries, measurement, and transformations: simple GIS analysis functions that, if clearly understood, can be strung together into powerful algorithms capable of solving sophisticated geographic problems § Much more to learn on this subject: stay tuned for GEOG 482! GEOG 380 - Topic 11 31