title: ’Análisis Estadístico: Accident Longitude

database <- read.csv("database-_1_.csv", header = TRUE, sep = ",", dec = ".", check.names = FALSE)
variable_original <- database$`Accident Longitude`
variable_original <- as.numeric(as.character(variable_original))
variable_original <- na.omit(variable_original)

amperaje <- variable_original

cat("Datos cargados correctamente. Total:", length(amperaje))

## Datos cargados correctamente. Total: 2795

cat("\nMínimo:", min(amperaje), "| Máximo:", max(amperaje))

## 
## Mínimo: -158.0999 | Máximo: 104.2634

# 1. Regla de Sturges
k <- 1 + (3.322 * log10(length(amperaje)))
k <- floor(k)

# 2. Rango y Amplitud
min_val <- min(amperaje)
max_val <- max(amperaje)
R <- max_val - min_val
A <- R / k

# 3. Límites de Intervalos
Li <- seq(from = min_val, to = max_val - A, by = A)
# Ajustes para asegurar cobertura
if(length(Li) < k) { Li <- c(Li, Li[length(Li)] + A) }
if(max(Li) + A < max_val) { Li <- c(Li, tail(Li, 1) + A) }

Ls <- Li + A
MC <- (Li + Ls) / 2

# 4. Frecuencias (ni)
ni <- numeric(length(Li))
for (i in 1:length(Li)) {
  if (i == length(Li)) {
     ni[i] <- sum(amperaje >= Li[i] & amperaje <= (max_val + 0.0001))
  } else {
     ni[i] <- sum(amperaje >= Li[i] & amperaje < Ls[i]) 
  }
}

# 5. Resto de columnas
hi <- ni / sum(ni) * 100
Niasc <- cumsum(ni)
Nidsc <- rev(cumsum(rev(ni)))
Hiasc <- round(cumsum(hi), 2)
Hidsc <- round(rev(cumsum(rev(hi))), 2)

TDFAmperaje <- data.frame(Li, Ls, MC, ni, hi, Niasc, Nidsc, Hiasc, Hidsc)
tabla1_sturges <- TDFAmperaje %>%
  gt() %>%
  tab_header(
    title = md("*Tabla 1: Distribución de Frecuencias*"),
    subtitle = md("**Variable: Accident Longitude**")
  ) %>%
  cols_label(
    Li = "L. Inferior",
    Ls = "L. Superior",
    MC = "Marca Clase",
    ni = "Frec. Abs.",
    hi = "Frec. Rel. %",
    Niasc = "Ni Asc.",
    Nidsc = "Ni Desc.",
    Hiasc = "Hi Asc. %",
    Hidsc = "Hi Desc. %"
  ) %>%
  fmt_number(columns = c(Li, Ls, MC), decimals = 4) %>%
  fmt_number(columns = c(hi, Hiasc, Hidsc), decimals = 2, pattern = "{x}%")

tabla1_sturges

L. Inferior	L. Superior	Marca Clase	Frec. Abs.	Frec. Rel. %	Ni Asc.	Ni Desc.	Hi Asc. %	Hi Desc. %
Tabla 1: Distribución de Frecuencias
Variable: Accident Longitude
−158.0999	−136.2363	−147.1681	14	0.50%	14	2795	0.50%	100.00%
−136.2363	−114.3727	−125.3045	169	6.05%	183	2781	6.55%	99.50%
−114.3727	−92.5091	−103.4409	1814	64.90%	1997	2612	71.45%	93.45%
−92.5091	−70.6455	−81.5773	792	28.34%	2789	798	99.79%	28.55%
−70.6455	−48.7819	−59.7137	2	0.07%	2791	6	99.86%	0.21%
−48.7819	−26.9183	−37.8501	0	0.00%	2791	4	99.86%	0.14%
−26.9183	−5.0547	−15.9865	1	0.04%	2792	4	99.89%	0.14%
−5.0547	16.8090	5.8772	0	0.00%	2792	3	99.89%	0.11%
16.8090	38.6726	27.7408	0	0.00%	2792	3	99.89%	0.11%
38.6726	60.5362	49.6044	0	0.00%	2792	3	99.89%	0.11%
60.5362	82.3998	71.4680	0	0.00%	2792	3	99.89%	0.11%
82.3998	104.2634	93.3316	3	0.11%	2795	3	100.00%	0.11%

total_ni <- sum(TDFAmperaje$ni)
total_hi <- sum(TDFAmperaje$hi)

TDF_Completo <- rbind(
  TDFAmperaje, 
  data.frame(
    Li = 0, Ls = 0, MC = 0,
    ni = total_ni, hi = total_hi, 
    Niasc = 0, Nidsc = 0, Hiasc = 0, Hidsc = 0
  )
)

tabla2_total <- TDF_Completo %>%
  gt() %>%
  tab_header(
    title = md("*Tabla 2: Distribución Completa*"),
    subtitle = md("**Con Totales**")
  ) %>%
  text_transform(
    locations = cells_body(columns = Li, rows = nrow(TDF_Completo)),
    fn = function(x) "Total"
  ) %>%
  text_transform(
    locations = cells_body(columns = c(Ls, MC, Niasc, Nidsc, Hiasc, Hidsc), rows = nrow(TDF_Completo)),
    fn = function(x) "" 
  ) %>%
  fmt_number(columns = c(Li), decimals = 0) # Truco visual para el texto Total

tabla2_total

Li	Ls	MC	ni	hi	Niasc	Nidsc	Hiasc	Hidsc
Tabla 2: Distribución Completa
Con Totales
−158	-136.236319	-147.168125	14	0.50089445	14	2795	0.50	100.00
−136	-114.372709	-125.304514	169	6.04651163	183	2781	6.55	99.50
−114	-92.509098	-103.440903	1814	64.90161002	1997	2612	71.45	93.45
−93	-70.645487	-81.577292	792	28.33631485	2789	798	99.79	28.55
−71	-48.781876	-59.713682	2	0.07155635	2791	6	99.86	0.21
−49	-26.918265	-37.850071	0	0.00000000	2791	4	99.86	0.14
−27	-5.054655	-15.986460	1	0.03577818	2792	4	99.89	0.14
−5	16.808956	5.877151	0	0.00000000	2792	3	99.89	0.11
17	38.672567	27.740761	0	0.00000000	2792	3	99.89	0.11
39	60.536178	49.604372	0	0.00000000	2792	3	99.89	0.11
61	82.399788	71.467983	0	0.00000000	2792	3	99.89	0.11
82	104.263399	93.331594	3	0.10733453	2795	3	100.00	0.11
Total			2795	100.00000000

hist(amperaje, 
     main = "Distribución de Longitud (Automático)",
     xlab = "Longitud", col = "lightgreen", las=1)

hist(amperaje, 
     breaks = c(Li, tail(Ls, 1)), 
     main = "Frecuencia (Sturges)",
     col = "darkgreen", 
     xlab = "Longitud", ylab = "Frecuencia")

boxplot(amperaje, 
        horizontal = TRUE, 
        col = "lightgreen", 
        main = "Boxplot: Accident Longitude")

title: ’Análisis Estadístico: Accident Longitude

Joham

2025-12-11