Introduction
These functions power the “Analyze Events” pane in interactShiny.
The idea is that each agent in a two-person interaction is an
independent InteRactModel.
situation <- define_situation(
agent1 = interact(), ## the idea is that each agent can be set up with
agent2 = interact() ## different dictionaries and equations.
)
#> ✔ dictionary = list(dataset = "usfullsurveyor2015", group = "all")
#> ✔ equations = list(key = "us2010", group = "all")
#> ✔ dictionary = list(dataset = "usfullsurveyor2015", group = "all")
#> ✔ equations = list(key = "us2010", group = "all")The newly created situation object contains various
fields and functions.
Most of these will be empty until the $start() function
is used.
For example:
situation$active ## one of the two agents will be active at any given time
#> [1] "agent1"
situation$time
#> NULL
situation$history
#> NULLAnd here is how they look after $start():
situation$start(list(A = "boyfriend", B = "hurt", O = "firefighter"))
opt_beh <- situation$optimal_behavior(who = "agent2")
opt_beh
#> # A tibble: 1 × 3
#> Be Bp Ba
#> <dbl> <dbl> <dbl>
#> 1 1.40 4.30 0.0535
situation$agent2$closest_terms(opt_beh, component = "behavior")[1:5]
#> [1] NA NA NA NA NA
situation$activate("agent2")
situation$new(list(A = "firefighter", B = "babble_to", O = "boyfriend"))
situation$history
#> $deflection
#> # A tibble: 2 × 7
#> time agent1 agent2 A B O deflection
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl>
#> 1 0 A O boyfriend hurt firefighter 98.7
#> 2 1 O A firefighter babble_to boyfriend 72.4
#>
#> $fundamentals
#> # A tibble: 2 × 15
#> time agent1 agent2 A B O Ae Ap Aa Be Bp Ba
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 A O boyfriend hurt fire… 1.86 1.68 0.89 -3.17 1.06 0.92
#> 2 1 O A firefight… babb… boyf… 3.27 2.85 2.29 -0.89 -0.66 1.88
#> # ℹ 3 more variables: Oe <dbl>, Op <dbl>, Oa <dbl>
#>
#> $transients
#> # A tibble: 2 × 15
#> time agent1 agent2 A B O Ae Ap Aa Be Bp Ba
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 A O boyfri… hurt fire… -3.58 1.90 1.67 -5.13 2.32 1.77
#> 2 1 O A firefi… babb… boyf… -0.423 -1.25 0.794 -0.652 -0.755 1.69
#> # ℹ 3 more variables: Oe <dbl>, Op <dbl>, Oa <dbl>
#>
#> $element_wise_deflection
#> # A tibble: 2 × 15
#> time agent1 agent2 A B O Ae Ap Aa Be Bp
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 A O boyfriend hurt fire… 29.6 0.0502 0.609 3.84 1.60
#> 2 1 O A firefighter babb… boyf… 13.6 16.8 2.24 0.0566 0.00894
#> # ℹ 4 more variables: Ba <dbl>, Oe <dbl>, Op <dbl>, Oa <dbl>This situation allows for “method chaining” using the
$ operator. This type of code is not very R-like.
situation$
activate("agent1")$
start(list(A = "employer", B = "hurt", O = "employee"))$
activate("agent2")$
new(list(A = "employee", B = "pull_away_from", O = "employer"))$
activate("agent1")$
new(list(A = "employer", B = "confront", O = "employee"))$
history
#> $deflection
#> # A tibble: 3 × 7
#> time agent1 agent2 A B O deflection
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl>
#> 1 0 A O employer hurt employee 43.5
#> 2 1 O A employee pull_away_from employer 19.2
#> 3 2 A O employer confront employee 13.3
#>
#> $fundamentals
#> # A tibble: 3 × 15
#> time agent1 agent2 A B O Ae Ap Aa Be Bp Ba
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 A O employer hurt empl… 1.57 2.56 0.7 -3.17 1.06 0.92
#> 2 1 O A employee pull_a… empl… 1.04 -0.37 0.2 -0.81 0.93 -0.6
#> 3 2 A O employer confro… empl… 1.57 2.56 0.7 0.39 1.99 1.28
#> # ℹ 3 more variables: Oe <dbl>, Op <dbl>, Oa <dbl>
#>
#> $transients
#> # A tibble: 3 × 15
#> time agent1 agent2 A B O Ae Ap Aa Be Bp Ba
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 A O emplo… hurt empl… -1.79 3.05 1.62 -2.38 2.79 1.55
#> 2 1 O A emplo… pull… empl… -0.486 -1.26 -1.33 -0.472 -0.449 -0.680
#> 3 2 A O emplo… conf… empl… -0.662 1.83 1.35 -0.607 1.76 1.28
#> # ℹ 3 more variables: Oe <dbl>, Op <dbl>, Oa <dbl>
#>
#> $element_wise_deflection
#> # A tibble: 3 × 15
#> time agent1 agent2 A B O Ae Ap Aa Be Bp Ba
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 A O employ… hurt empl… 11.3 0.240 0.848 0.622 3.00 3.98e-1
#> 2 1 O A employ… pull… empl… 2.33 0.795 2.33 0.114 1.90 6.43e-3
#> 3 2 A O employ… conf… empl… 4.98 0.531 0.424 0.995 0.0511 1.94e-5
#> # ℹ 3 more variables: Oe <dbl>, Op <dbl>, Oa <dbl>
situation$history$deflection
#> # A tibble: 3 × 7
#> time agent1 agent2 A B O deflection
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl>
#> 1 0 A O employer hurt employee 43.5
#> 2 1 O A employee pull_away_from employer 19.2
#> 3 2 A O employer confront employee 13.3Instead, you can use R’s pipe operator in tandem with the
sttn_* functions.
df <- situation |>
sttn_activate("agent1") |>
sttn_start(list(A = "employer", B = "hurt", O = "employee")) |>
sttn_activate("agent2") |>
sttn_new(list(A = "employee", B = "pull_away_from", O = "employer")) |>
sttn_activate("agent1") |>
sttn_new(list(A = "employer", B = "confront", O = "employee")) |>
sttn_extract()
df$deflection
#> # A tibble: 3 × 7
#> time agent1 agent2 A B O deflection
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl>
#> 1 0 A O employer hurt employee 43.5
#> 2 1 O A employee pull_away_from employer 19.2
#> 3 2 A O employer confront employee 13.3
df$transients
#> # A tibble: 3 × 15
#> time agent1 agent2 A B O Ae Ap Aa Be Bp Ba
#> <int> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0 A O emplo… hurt empl… -1.79 3.05 1.62 -2.38 2.79 1.55
#> 2 1 O A emplo… pull… empl… -0.486 -1.26 -1.33 -0.472 -0.449 -0.680
#> 3 2 A O emplo… conf… empl… -0.662 1.83 1.35 -0.607 1.76 1.28
#> # ℹ 3 more variables: Oe <dbl>, Op <dbl>, Oa <dbl>