Skip to main content

Accident count by weather condition

Importing the package

library(amazonasdatahub)

Accident count by weather condition​

With the prf_amazonas database, we can perform data analysis of accidents recorded by the Federal Highway Police (PRF) in the state of Amazonas in the year 2025.

We can check the frequency of accidents by weather condition and identify which of these conditions presents a higher frequency of accidents.

# Importanto pacotes para organizacao e visualizacao de dados
require(dplyr)
require(forcats)
require(ggplot2)

# Contando acidentes agrupados por condicao climatica e fazendo o grafico
prf_amazonas %>%
  count(weather) %>%
  mutate(
    weather = fct_reorder(weather, -n)
  ) %>%
  ggplot(aes(x = weather, y = n)) +
  geom_bar(stat = "identity", fill = "steelblue") +
  geom_text(aes(label = n), vjust = -0.5) +
  theme_minimal() +
  labs(
    title = "Contagem de acidentes por condição climática",
    x = "Condição climática",
    y = "Contagem"
  )

See that the Céu claro condition is divided, since one has written with an acute accent Céu claro and another is without an acute accent, Ceu claro. We can individually assess weather conditions available to check if it makes sense.

unique(prf_amazonas$weather)
 [1] "Ceu Claro"        "Nublado"          "Ignorada"         "Chuva"           
 [5] "Nevoeiro/neblina" "Sol"              "Vento"            "Céu Claro"       
 [9] "Ignorado"         "Nevoeiro/Neblina" "Garoa/Chuvisco"

Note that in addition to Céu claro, we have Nevoeiro/Neblina and Nevoeiro/neblina. This small change of capital letters already interferes with data reading. So let's standardize

prf_amazonas_clean <- prf_amazonas %>%
  mutate(
    weather = case_when(
      weather == 'Ceu Claro' ~ 'Céu Claro',
      weather == 'Nevoeiro/neblina' ~ 'Nevoeiro/Neblina',
      TRUE~weather
    )
  )

It is now possible to view the bar graph of the corrected data.

prf_amazonas_clean %>%
  count(weather) %>%
  mutate(
    weather = fct_reorder(weather, -n)
  ) %>%
  ggplot(aes(x = weather, y = n)) +
  geom_bar(stat = "identity", fill = "steelblue") +
  geom_text(aes(label = n), vjust = -0.5) +
  theme_minimal() +
  labs(
    title = "Contagem de acidentes por condição climática",
    x = "Condição climática",
    y = "Contagem"
  )