geomultistar: Multidimensional Queries Enriched with Geographic Data

Travis build status

Multidimensional systems allow complex queries to be carried out in an easy way. The geographical dimension, together with the temporal dimension, plays a fundamental role in multidimensional systems. Through the geomultistar package, vector layers can be associated to the attributes of geographic dimensions, so that the results of multidimensional queries can be obtained directly as vector layers. In other words, this package allows enriching multidimensional queries with geographic data.

The multidimensional structures on which we can define the queries can be created from flat tables with starschemar package or imported directly using functions from geomultistar package.

Installation

You can install the released version of geomultistar from CRAN with:

install.packages("geomultistar")

And the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("josesamos/geomultistar")

Example

If we start from a flat table, we can generate a star schema using the starschemar package, as described in its examples.

If we have a star schema in another tool, we need to import the fact and dimension tables into R in the form of tables implemented by tibble (mrs_fact_age, mrs_fact_cause, mrs_where, mrs_when and mrs_who in the example). Once we have them in this format, we have to build a multistar structure from them: This structure can contain multiple fact and dimension tables, so facts can share dimensions. The definition for tables obtained from the case detailed in starschemar is included below. The measures of the facts are defined and the relationships between facts and dimensions are established.

library(tidyr)
library(geomultistar)

ms <- multistar() %>%
  add_facts(
    fact_name = "mrs_age",
    fact_table = mrs_fact_age,
    measures = "n_deaths",
    nrow_agg = "count"
  ) %>%
  add_facts(
    fact_name = "mrs_cause",
    fact_table = mrs_fact_cause,
    measures = c("pneumonia_and_influenza_deaths", "other_deaths"),
    nrow_agg = "nrow_agg"
  ) %>%
  add_dimension(
    dimension_name = "where",
    dimension_table = mrs_where,
    dimension_key = "where_pk",
    fact_name = "mrs_age",
    fact_key = "where_fk"
  ) %>%
  add_dimension(
    dimension_name = "when",
    dimension_table = mrs_when,
    dimension_key = "when_pk",
    fact_name = "mrs_age",
    fact_key = "when_fk",
    key_as_data = TRUE
  ) %>%
  add_dimension(
    dimension_name = "who",
    dimension_table = mrs_who,
    dimension_key = "who_pk",
    fact_name = "mrs_age",
    fact_key = "who_fk"
  ) %>%
  relate_dimension(dimension_name = "where",
                   fact_name = "mrs_cause",
                   fact_key = "where_fk") %>%
  relate_dimension(dimension_name = "when",
                   fact_name = "mrs_cause",
                   fact_key = "when_fk")

Once we have a multistar structure, we will associate vector layers to the attributes of the geographic dimensions. We can use existing layers or generate them from the previous definitions. As a result we will have a geomultistar structure.

library(sf)
#> Linking to GEOS 3.8.0, GDAL 3.0.4, PROJ 6.3.1

gms <-
  geomultistar(ms, geodimension = "where") %>%
  define_geoattribute(
    attribute = "city",
    from_layer = usa_cities,
    by = c("city" = "city", "state" = "state")
  ) %>%
  define_geoattribute(
    attribute = "county",
    from_layer = usa_counties,
    by = c("county" = "county", "state" = "state")
  )  %>%
  define_geoattribute(
    attribute = c("state"),
    from_layer = usa_states,
    by = c("state" = "state")
  ) %>%
  define_geoattribute(from_attribute = "state")

In the last definition, because no geographic attribute is specified, the rest of the dimension’s attributes are automatically defined from the layer associated with the indicated parameter.

Finally, we can define multidimensional queries on this structure using the functions available in the starschemar package. When executing these queries, using the functionality implemented in package geomultistar, the vector layers of the attributes will be taken into account to result in a new vector layer.

library(starschemar)

gdqr <- dimensional_query(gms) %>%
  select_dimension(name = "where",
                   attributes = c("division_name", "region_name")) %>%
  select_dimension(name = "when",
                   attributes = c("year", "week")) %>%
  select_fact(name = "mrs_age",
              measures = c("n_deaths")) %>%
  select_fact(
    name = "mrs_cause",
    measures = c("pneumonia_and_influenza_deaths", "other_deaths")
  ) %>%
  filter_dimension(name = "when", week <= "03") %>%
  run_geoquery(wider = TRUE)

The result is a vector layer that we can save or we can see it as a map, using the functions associated with the sf class.

class(gdqr)
#> [1] "list"

plot(gdqr$sf[,"n_deaths_01"])

Although we have indicated in the query the attributes division_name and region_name, as can be seen in the figure, the result obtained is at the finest granularity level, in this case at the division_name level.

Only the parts of the divisions made up of states where there is recorded data are shown. If we wanted to show the full extent of each division, we should have explicitly associated a layer at that level.

The result includes the meaning of each variable in table form.

id_variable measure week
n_deaths_01 n_deaths 01
n_deaths_02 n_deaths 02
n_deaths_03 n_deaths 03
count_01 count 01
count_02 count 02
count_03 count 03
pneumonia_and_influenza_deaths_01 pneumonia_and_influenza_deaths 01
pneumonia_and_influenza_deaths_02 pneumonia_and_influenza_deaths 02
pneumonia_and_influenza_deaths_03 pneumonia_and_influenza_deaths 03
other_deaths_01 other_deaths 01
other_deaths_02 other_deaths 02
other_deaths_03 other_deaths 03

It can be saved directly as a GeoPackage, using the save_as_geopackage function.

save_as_geopackage(vl_sf_w, "division")