The goal of SchellingR is to run the Schelling (1971) agent-based model (ABM) of urban migration and segregation. The ABM is written in C++ to allow for fast and efficient calculations. Summary statistics of each round are reported back to R.
You can install SchellingR from github with:
# install.packages("devtools")
devtools::install_github("DavZim/SchellingR")The base function of the package is run_schelling() which returns a
list of two data.frames round and detailed. The former contains
aggregated statistics per round, i.e., the number of happy/unhappy
agents and the number of moves, the latter contains agent-level
information about the grid, the agents, and their neighbors.
The function allows you to tweak basic inputs to the model, such as grid
size size, the percent of empty grid-cells percent_empty, the
threshold above which agents are considered to be happy threshold, the
number of groups to simulated number_of_groups, the maximum number of
simulation rounds max_rounds, and a seed for the random number
generator seed.
library(SchellingR)
set.seed(1234567)
sh <- run_schelling(size = 10, percent_empty = 0.2, threshold = 0.5, number_of_groups = 2, max_rounds = 100)
str(sh)
#> List of 2
#> $ round :Classes 'tbl_df', 'tbl' and 'data.frame': 11 obs. of 4 variables:
#> ..$ round : int [1:11] 0 1 2 3 4 5 6 7 8 9 ...
#> ..$ happy_agents : int [1:11] 42 54 68 73 75 76 77 78 79 79 ...
#> ..$ unhappy_agents: int [1:11] 38 26 12 7 5 4 3 2 1 1 ...
#> ..$ number_moves : int [1:11] 0 38 26 12 7 5 4 3 2 1 ...
#> $ detailed:Classes 'tbl_df', 'tbl' and 'data.frame': 1100 obs. of 8 variables:
#> ..$ round : int [1:1100] 0 0 0 0 0 0 0 0 0 0 ...
#> ..$ x : int [1:1100] 1 2 3 4 5 6 7 8 9 10 ...
#> ..$ y : int [1:1100] 10 10 10 10 10 10 10 10 10 10 ...
#> ..$ id : int [1:1100] 0 1 2 3 4 5 6 7 8 0 ...
#> ..$ group : int [1:1100] 2 1 2 1 2 1 2 1 1 0 ...
#> ..$ n_same : int [1:1100] 0 2 2 1 1 1 2 2 3 0 ...
#> ..$ n_different: int [1:1100] 3 3 3 4 3 3 2 3 1 0 ...
#> ..$ happy : int [1:1100] 0 0 0 0 0 0 1 0 1 0 ...For full functionality, you need to have the additional packages installed
gganimatefor animation (gifs)emojifontfor emojismagickfor combining gifs
After running the simulation, we can plot the average happiness per
round using dplyr and ggplot2.
plot_development(sh)
We can also use the detailed information to plot the grid at certain steps.
plot_grid(sh, select_round = 8, title = TRUE)
Using the wonderful gganimate package, we can also create an animation of the model
plot_grid(sh, title = TRUE, animate = TRUE)
If you want to get even fancier, you can also use the emojifont package to make the agent’s state more intuitive.
plot_grid(sh, select_round = 8, title = TRUE, emoji = TRUE)
Or you can use both options at the same time.
plot_grid(sh, title = TRUE, emoji = TRUE, animate = TRUE)
To give a more complex example on a larger grid with 7 groups:
sh2 <- run_schelling(size = 50, number_of_groups = 5, max_rounds = 1000, seed = 1234567)
grid <- plot_grid(sh2, title = TRUE, animate = TRUE, transition = FALSE)
dev <- plot_development(sh2, animate = TRUE)
combined <- combine_gifs(grid, dev)
#> geom_path: Each group consists of only one observation. Do you need to adjust
#> the group aesthetic?
#> geom_path: Each group consists of only one observation. Do you need to adjust
#> the group aesthetic?
print(combined)
#> # A tibble: 100 x 7
#> format width height colorspace matte filesize density
#> <chr> <int> <int> <chr> <lgl> <int> <chr>
#> 1 GIF 1344 480 sRGB TRUE 0 72x72
#> 2 GIF 1344 480 sRGB TRUE 0 72x72
#> 3 GIF 1344 480 sRGB TRUE 0 72x72
#> 4 GIF 1344 480 sRGB TRUE 0 72x72
#> 5 GIF 1344 480 sRGB TRUE 0 72x72
#> 6 GIF 1344 480 sRGB TRUE 0 72x72
#> 7 GIF 1344 480 sRGB TRUE 0 72x72
#> 8 GIF 1344 480 sRGB TRUE 0 72x72
#> 9 GIF 1344 480 sRGB TRUE 0 72x72
#> 10 GIF 1344 480 sRGB TRUE 0 72x72
#> # … with 90 more rows