Jingwen Zheng
  • Apr 06, 2017

R: Working with GeoJSON data and resolve 'Error in ogrInfo'

This blog presents how to work with GeoJSON data in R, and proposes the solution of "Error in ogrInfo(dsn = dsn, layer = layer, encoding = encoding, use_iconv = use_iconv"

Since working in franprix, I’ve accomplished many projects, for different departments. Thanks to this opportunity, I understood business of various services and learnt lots of technique of R. Today I’d like to share with you how to read GeoJSON data in R.

Concerning the dataset, we take the Paris international tourist area as an example. Then we use either rgdal or geojsonio package, and read it as following.

ztiData <- "zones-touristiques-internationales-paris.geojson"
ztiPolygon <- rgdal::readOGR(dsn = ztiData,
                             layer = "OGRGeoJSON",
                             require_geomType = "wkbPolygon",
                             encoding = "UTF-8")
# or use the geojson_read() function
# ztiPolygon <- geojsonio::geojson_read(ztiData, what = "sp")

# then make a map with following scripts
ztiParis <- leaflet(ztiPolygon) %>%
            addTiles() %>%
            addPolygons(stroke = FALSE,
                        smoothFactor = 0.3,
                        fillOpacity = 0.5,
                        fillColor = "blue",
                        label = ~ paste0(name))

Moreover, for reading GeoJSON data, many web sites recommend similar codes but without require_geomType parameter, and many people get following errors:

ztiPolygon <- rgdal::readOGR(dsn = ztiData, layer = "OGRGeoJSON")

> Error in ogrInfo(dsn = dsn, layer = layer, encoding = encoding,
> use_iconv = use_iconv,  :
> Multiple incompatible geometries: wkbPoint: 6; wkbPolygon: 13

Thanks to stackoverflow, I understood what the error means, and extract the features that I need.

ogrListLayers(ztiData)
ogrInfo(ztiData, "OGRGeoJSON")
ogrInfo(ztiData, "OGRGeoJSON", require_geomType="wkbPoint")
# NOTE: keeping only 6 wkbPoint of 19 features
# note that extent applies to all features
# Driver: GeoJSON; number of rows: 19 
# selected geometry type: wkbPoint with 6 rows
# Feature type: wkbPoint with 2 dimensions
# Extent: (2.280886 48.82001) - (2.388132 48.88772)
# CRS: +proj=longlat +datum=WGS84 +no_defs  
# Number of fields: 17
#
#                     name type length typeName
# 1         nom_de_la_gare    4      0   String
# 2                adresse    4      0   String
# ...

ogrInfo(ztiData, "OGRGeoJSON", require_geomType="wkbPolygon")
# NOTE: keeping only 13 wkbPolygon of 19 features
# note that extent applies to all features
# Driver: GeoJSON; number of rows: 19 
# selected geometry type: wkbPolygon with 13 rows
# Feature type: wkbPolygon with 2 dimensions
# Extent: (2.280886 48.82001) - (2.388132 48.88772)
# CRS: +proj=longlat +datum=WGS84 +no_defs  
# Number of fields: 17

points <- readOGR(ztiData,
                  "OGRGeoJSON",
                  require_geomType="wkbPoint")
# Selected wkbPoint feature type, with 6 rows
# It has 17 fields
# NOTE: keeping only 6 wkbPoint of 19 features

polygons <- readOGR(ztiData,
                    "OGRGeoJSON",
                    require_geomType="wkbPolygon")
# Selected wkbPolygon feature type, with 13 rows
# It has 17 fields
# NOTE: keeping only 13 wkbPolygon of 19 features

# prove they red and orange in things
plot(points, add = TRUE, col = "red")
plot(polygons, add = TRUE, col = "orange")

After confirming with codes above, I could make sure that wkbPolygon is the parameter needed, so it was added in the codes at the beginning.

This method can be also applied with the result of leaflet(). We can plot some places in the map with leaflet(), then use addPolygons() to add the ploygons-layer.

Search