In Class exercise 3

3.0 Notes:

• Conduct a Monte Carlo simulation to test for Complete Spatial Randomness (CSR)

• Monte Carlo simulation test of CSR:

  • To determine the simulation envelope, use the 95th percentile (maximum) and 5th percentile (minimum) values of G(r) from the simulations.

  • When randomising data it is important to set seed to have it be repeatable

• Nearest Neighbour Index:

  • Before hypothesis, one should determine confidence interval (confidence level) to justify conclusion reached

  • Due to the unpredictability of real life data, uncertainty would need to be considered via the confidence level

    • 99 - 99.9 confidence level should be avoided due to the perceived notion that it is almost fully accurate, which cannot happen due to real world uncertainty

  • Reject P-value, if P-value< Alpha value

• L Functions Interpretation:

  • Signs of clustering can be determined from how much higher the L value is above the envelope

• Ripley’s K function:

  • Both G function and K function are distance based, but G function is for any particular zone (isolated), but K function is cumulative in nature (inclusive)

  • Usage lies in zoning based on the various interval ranges

3.1 Getting Started

Maptools is retired and binary is removed from CRAN. However, we can download froom Posit Public Package Manager

install.packages("maptools", repos = "https://packagemanager.posit.co/cran/2023-10-13")

pacman::p_load(sf, raster, spatstat, tmap, tidyverse)

3.1.1 Working with st_union()

The code chunk below is used to derive the coastal outline in tibble data frame sg_sf <- mpsz_sf %>% st_union()

3.2 Viewing data for Take Home exercise 1

acled_sf <- read_csv("data/ACLED_Myanmar.csv") %>% 
  st_as_sf(coords = c(
    "longitude", "latitude"), crs = 4326) %>% 
  st_transform(crs= 32647) %>%
  mutate(event_date = dmy(event_date))

tmap_mode('view')
acled_sf %>%
  filter(year == 2023 |
           event_type == "Political violence") %>%
  tm_shape()+
  tm_dots()
tmap_mode("plot")