class: left, top, title-slide # Geospatial Data<br>Geospatial Mapping ### Keith VanderLinden<br>Calvin University --- # Geospatial Datasets .pull-left[ `spData::world` is an R list that represents an simple-feature (`sf`) collection including: - Bounding box - Coordinate Reference System - Data for each country, including vector geometry specifications encoded as multi-polygons ```r library(tidyverse) library(sf) library(spData) ``` ] .pull-right[ ```r spData::world ``` ``` Simple feature collection with 177 features and 10 fields Geometry type: MULTIPOLYGON Dimension: XY Bounding box: xmin: -180 ymin: -89.9 xmax: 180 ymax: 83.64513 Geodetic CRS: WGS 84 # A tibble: 177 x 11 iso_a2 name_long continent region_un subregion type area_km2 * <chr> <chr> <chr> <chr> <chr> <chr> <dbl> 1 FJ Fiji Oceania Oceania Melanesia Sove~ 1.93e4 2 TZ Tanzania Africa Africa Eastern ~ Sove~ 9.33e5 3 EH Western Sa~ Africa Africa Northern~ Inde~ 9.63e4 4 CA Canada North Am~ Americas Northern~ Sove~ 1.00e7 5 US United Sta~ North Am~ Americas Northern~ Coun~ 9.51e6 6 KZ Kazakhstan Asia Asia Central ~ Sove~ 2.73e6 # ... with 171 more rows, and 4 more variables: pop <dbl>, # lifeExp <dbl>, gdpPercap <dbl>, geom <MULTIPOLYGON [°]> ``` ] ??? - `spData::world` is an R list that represents an simple-feature (`sf`) collection of 177 countries, 10 fields per country. - Bounding box is most of the globe: -180–180 logitude and -89.99–83.59 latitude - CRS: the WGS84 coordinate reference system (see slide below) - Cf. the world data from `rnaturalearth`, which includes a dataset with 241 countries, 63 fields per country. --- # Mapping Geospatial Data: The World `spData::world` contains geometric data on all countries. Such data are most compelling when presented in the form of a *map*. .pull-left[ `ggplot` & `sf` work together to plot this data easily. ```r spData::world %>% ggplot() + geom_sf() ``` ] .pull-right[ <img src="mapping_files/figure-html/unnamed-chunk-3-1.png" width="100%" style="display: block; margin: auto;" /> ] ??? --- # Mapping Geospatial Data: One Country .pull-left[ ```r spData::world$name_long[2] ``` ``` [1] "Tanzania" ``` ```r spData::world$geom[2] ``` ``` Geometry set for 1 feature Geometry type: MULTIPOLYGON Dimension: XY Bounding box: xmin: 29.34 ymin: -11.72094 xmax: 40.31659 ymax: -0.95 Geodetic CRS: WGS 84 ``` ``` MULTIPOLYGON (((33.90371 -0.95, 31.86617 -1.027... ``` ] .pull-right[ ```r world$geom[2] %>% ggplot() + geom_sf() + geom_sf_text( aes(label = world$name_long[2])) ``` <img src="mapping_files/figure-html/unnamed-chunk-5-1.png" width="100%" style="display: block; margin: auto;" /> ] ??? - Here we focus on the 2nd country in the dataset: Tanzania. - The coordinates are x (longitude); y (latitude). - The geometry is encoded using *Well-Known Text* (WKT) format, which supports points, lines, polygons, and multi-* versions of all three. (N.b., I had to cut-paste the geometry text to get it onto the slide; it wouldn't display otherwise.) --- # Coordinate Reference Systems A *Coordinate Reference System* (CRS) specifies locations on the surface of the Earth using angular *coordinates*. We focus on *geoditic* coordinate systems, which use *latitude* and *longitude*. .pull-left[  ] .pull-right[ Common CRS systems include: - [WGS84](http://wiki.gis.com/wiki/index.php/WGS84) (aka EPSG4326): Used by GPS - [NAD83](http://wiki.gis.com/wiki/index.php/NAD83): North American system - OSGB: UK standard used for the Cholera data, which created issues discussed in the text ] .footnote[Image from: [wikimedia.org](https://en.wikipedia.org/wiki/Geographic_coordinate_system)] ??? - Geodetic CRSs use angular measurements. - Latitude: -90 – +90 (N.b., That's 90 degrees from equator to pole.) - Longitude: -180 – +180 (N.b., That's 360 degrees around the equitorial circle) - References: - https://geocompr.robinlovelace.net/spatial-class.html#crs-intro - https://datacarpentry.org/organization-geospatial/03-crs/index.html - https://www.esri.com/arcgis-blog/products/arcgis-pro/education/earth-peel/ - https://www.geo-projections.com/ --- # Projections *Projections* map CRS coordinates onto planar coordinates, projecting them onto either a cylinder, cone, or plane. This preserves Shape/Angle (*conformal*) **XOR** Area (*equal-area*).  .footnote[Image from: [Map Projections](https://gistbok.ucgis.org/bok-topics/2017-quarter-03/map-projections)] ??? - A globe would be more accurate, but it can't be included in books or shown on computer screens. - x-y projections of lat-long coordinates are, necessarily, deformed in some way. (see: [Orange Peel Analogy](https://datacarpentry.org/organization-geospatial/03-crs)). - Choose a projection that suits your usecase (cf. MDSR 17.5: How (not) to lie). --- # Projections: Examples ```r library(mapproj) library(maps) ``` .pull-left[ **Mercator** ```r maps::map( "world", projection = "mercator", wrap = TRUE ) ``` <img src="mapping_files/figure-html/unnamed-chunk-7-1.png" width="90%" style="display: block; margin: auto;" /> ] .pull-right[ **Gall-Peters** ```r maps::map( "world", projection = "cylequalarea", param = 0, wrap = TRUE) ``` <img src="mapping_files/figure-html/unnamed-chunk-8-1.png" width="90%" style="display: block; margin: auto;" /> ] ??? - Libraries: - `mapproj` does map projections. - `maps` draws the maps. - mercator() - Equatorial projections centered on the Prime Meridian (longitude 0). Parallels are straight horizontal lines. - equally spaced straight meridians, **conformal**, straight compass courses - cylindrical - cylequalarea(lat0) - aka. Gall-Peters (cited in the text) - equally spaced straight meridians, **equal-area**, true scale on lat0 - **cylindrical** (centered on param = lat0) --- # Projections: Examples (continued) ```r library(mapproj) library(maps) ``` .pull-left[ **Lambert** ```r maps::map( "world", projection = "lambert", parameters = c(lat0 = 0, lat1 = 90), wrap = TRUE) ``` <img src="mapping_files/figure-html/unnamed-chunk-10-1.png" width="90%" style="display: block; margin: auto;" /> ] .pull-right[ **Albers** ```r maps::map( "world", projection = "albers", parameters = c(lat0 = 20, lat1 = 50), wrap = TRUE) ``` <img src="mapping_files/figure-html/unnamed-chunk-11-1.png" width="90%" style="display: block; margin: auto;" /> ] ??? - lambert(lat0,lat1) - **conformal**, true scale on lat0 and lat1 - **Conic** - The parameters specify where the cylider hits the Earth (here, the cone point is over Greenwich). - albers(lat0,lat1) - **equal-area**, true scale on lat0 and lat1 - **Conic** - Not great for the world, but decent for smaller regions (e.g., "state")