Norway’s Crime Revolution: How Reported Offences Collapsed Across Categories

SSB

crime

society

policing

Three decades of crime data reveal a dramatic transformation in Norwegian society — but not all offences are declining equally

Published

April 10, 2026

Norway has witnessed one of the most dramatic crime declines in the developed world, yet the story beneath the headline numbers reveals a complex transformation. While property crimes have plummeted, other categories tell different stories — some rising, others stubbornly persistent. This analysis examines three decades of police-reported offences to understand what’s really happening on Norwegian streets.

Setup

Code

knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = FALSE, error = TRUE)

library(tidyverse)
library(PxWebApiData)
library(lubridate)
library(scales)
library(MetBrewer)

pal <- met.brewer("Hokusai1", 7)

Data Collection

Code

df <- NULL

tryCatch({
  raw <- ApiData(
    "https://data.ssb.no/api/v0/no/table/08484",
    LovbruddKrim = TRUE,
    ContentsCode = TRUE,
    Tid = list(filter = "top", values = 32)
  )
  
  tmp <- raw[[1]]
  message("Columns: ", paste(names(tmp), collapse = ", "))
  message("Rows fetched: ", nrow(tmp))
  print(head(tmp, 10))
  
  # Detect time column
  time_col <- names(tmp)[grepl(
    "tid|\u00e5r|kvartal|m\u00e5ned|aar|maaned|year|month|quarter",
    names(tmp), ignore.case = TRUE, perl = TRUE
  )][1]
  if (is.na(time_col)) time_col <- names(tmp)[length(names(tmp)) - 1L]
  
  # Detect value column
  value_col <- names(tmp)[vapply(tmp, is.numeric, logical(1L))][1]
  if (is.na(value_col)) value_col <- names(tmp)[length(names(tmp))]
  
  # Detect crime type column
  crime_col <- names(tmp)[grepl("lovbrudd|offence|crime", names(tmp), ignore.case = TRUE)][1]
  if (is.na(crime_col)) stop("Cannot detect crime column — columns are: ", paste(names(tmp), collapse = ", "))
  
  # Detect contents/statistic type column
  contents_col <- names(tmp)[grepl("statistikkvariabel|contents|innhold|komponent", names(tmp), ignore.case = TRUE)][1]
  if (is.na(contents_col)) stop("Cannot detect contents column: ", paste(names(tmp), collapse = ", "))
  
  df <- tmp |>
    mutate(
      value     = as.numeric(.data[[value_col]]),
      time_str  = .data[[time_col]],
      crime_type = .data[[crime_col]],
      stat_type  = .data[[contents_col]],
      date      = case_when(
        stringr::str_detect(time_str, "M") ~ lubridate::ym(sub("M", "-", time_str)),
        stringr::str_detect(time_str, "K") ~ lubridate::yq(sub("K", " Q", time_str)),
        nchar(time_str) == 4               ~ lubridate::ymd(paste0(time_str, "-01-01")),
        TRUE ~ NA_Date_
      ),
      year = year(date)
    ) |>
    filter(!is.na(value), !is.na(date))
  
  message("Clean rows after filter: ", nrow(df))
  if (nrow(df) == 0) stop("Data frame is empty after cleaning")
  
}, error = function(e) {
  message("DATA FETCH FAILED: ", e$message)
  message("df will be NULL — no plots will render")
})

           lovbruddstype statistikkvariabel   år value NAstatus
1  Alle lovbruddsgrupper   Lovbrudd anmeldt 1993    NA        .
2  Alle lovbruddsgrupper   Lovbrudd anmeldt 1994    NA        .
3  Alle lovbruddsgrupper   Lovbrudd anmeldt 1995    NA        .
4  Alle lovbruddsgrupper   Lovbrudd anmeldt 1996    NA        .
5  Alle lovbruddsgrupper   Lovbrudd anmeldt 1997    NA        .
6  Alle lovbruddsgrupper   Lovbrudd anmeldt 1998    NA        .
7  Alle lovbruddsgrupper   Lovbrudd anmeldt 1999    NA        .
8  Alle lovbruddsgrupper   Lovbrudd anmeldt 2000    NA        .
9  Alle lovbruddsgrupper   Lovbrudd anmeldt 2001    NA        .
10 Alle lovbruddsgrupper   Lovbrudd anmeldt 2002    NA        .

The Total Picture: Three Decades of Change

Let’s start with the broadest view — total reported offences across all categories. The scale of the transformation becomes immediately clear.

Code

if (!is.null(df)) {
  
  total_df <- df |>
    filter(
      grepl("Alle lovbruddsgrupper|All", crime_type, ignore.case = TRUE),
      grepl("Lovbrudd anmeldt|reported", stat_type, ignore.case = TRUE)
    ) |>
    arrange(date)
  
  if (nrow(total_df) > 0) {
    peak_year <- total_df |> filter(value == max(value)) |> pull(year) |> first()
    peak_value <- max(total_df$value)
    recent_value <- total_df |> filter(year == max(year)) |> pull(value) |> first()
    
    p1 <- ggplot(total_df, aes(x = date, y = value)) +
      geom_area(fill = pal[1], alpha = 0.7) +
      geom_line(color = pal[7], linewidth = 1) +
      annotate("text", x = as.Date("2002-01-01"), y = peak_value - 10000,
               label = paste0("Peak: ", comma(peak_year), "\n", comma(round(peak_value/1000, 0)), "k offences"),
               family = "sans", size = 3.5, hjust = 0) +
      annotate("segment", x = as.Date("2004-01-01"), xend = as.Date(paste0(peak_year, "-01-01")),
               y = peak_value - 15000, yend = peak_value - 5000,
               arrow = arrow(length = unit(0.2, "cm")), color = pal[7]) +
      scale_y_continuous(labels = comma_format()) +
      labs(
        title = "Norway's Crime Collapse: Total Reported Offences Since 1993",
        subtitle = paste0("From ", comma(round(peak_value/1000, 0)), "k offences at peak to ", 
                         comma(round(recent_value/1000, 0)), "k today — a ", 
                         round((1 - recent_value/peak_value)*100, 0), "% decline"),
        caption = "Source: Statistics Norway (SSB) | Table 08484",
        x = NULL,
        y = "Reported offences"
      ) +
      theme_minimal(base_size = 12) +
      theme(
        plot.title = element_text(face = "bold", size = 14),
        plot.subtitle = element_text(color = "gray40", size = 11),
        panel.grid.minor = element_blank()
      )
    
    print(p1)
  }
}

Category Breakdown: Winners and Losers

Not all crime categories have followed the same trajectory. Let’s examine which categories have seen the largest absolute changes between the earliest and most recent years available.

Code

if (!is.null(df)) {
  
  category_df <- df |>
    filter(
      grepl("Lovbrudd anmeldt|reported", stat_type, ignore.case = TRUE),
      !grepl("Alle lovbruddsgrupper|All|total", crime_type, ignore.case = TRUE)
    ) |>
    group_by(crime_type) |>
    filter(n() >= 20) |>  # Only categories with substantial data
    arrange(date) |>
    mutate(
      first_value = first(value),
      last_value = last(value),
      change = last_value - first_value,
      pct_change = (last_value - first_value) / first_value * 100
    ) |>
    ungroup() |>
    distinct(crime_type, first_value, last_value, change, pct_change)
  
  if (nrow(category_df) > 0) {
    
    # Clean up crime type names - extract descriptive part
    category_df <- category_df |>
      mutate(
        crime_clean = str_remove(crime_type, "^[0-9A-Z]+AAAA-[0-9A-Z]+ZZZZz ¬+ "),
        crime_clean = str_remove(crime_clean, "^[0-9A-Z]+[A-Z][a-z]+ ¬+ "),
        crime_clean = str_trim(crime_clean)
      ) |>
      filter(nchar(crime_clean) > 3, abs(change) > 1000) |>
      arrange(change) |>
      slice(c(1:8, (n()-7):n()))  # Top and bottom 8
    
    p2 <- ggplot(category_df, aes(x = change, y = reorder(crime_clean, change))) +
      geom_segment(aes(x = 0, xend = change, y = crime_clean, yend = crime_clean),
                   color = "gray70", linewidth = 0.5) +
      geom_point(aes(color = change > 0), size = 4, show.legend = FALSE) +
      scale_color_manual(values = c("TRUE" = pal[6], "FALSE" = pal[1])) +
      scale_x_continuous(labels = comma_format()) +
      labs(
        title = "Norway's Crime Revolution by Category",
        subtitle = "Change in reported offences from 1993 to 2024 — property crimes collapsed while some categories persisted",
        caption = "Source: Statistics Norway (SSB) | Table 08484",
        x = "Change in reported offences",
        y = NULL
      ) +
      theme_minimal(base_size = 11) +
      theme(
        plot.title = element_text(face = "bold", size = 14),
        plot.subtitle = element_text(color = "gray40", size = 10),
        panel.grid.major.y = element_blank(),
        panel.grid.minor = element_blank()
      )
    
    print(p2)
  }
}

The Geographic Dimension: Per Capita Rates

Crime rates per 1,000 inhabitants tell a different story than raw numbers — accounting for population growth reveals whether society is actually becoming safer or just growing faster than crime.

Code

if (!is.null(df)) {
  
  per_capita_df <- df |>
    filter(
      grepl("Alle lovbruddsgrupper|All", crime_type, ignore.case = TRUE),
      grepl("per 1000|per capita", stat_type, ignore.case = TRUE)
    ) |>
    arrange(date)
  
  if (nrow(per_capita_df) > 0) {
    
    first_rate <- per_capita_df |> filter(year == min(year)) |> pull(value) |> first()
    last_rate <- per_capita_df |> filter(year == max(year)) |> pull(value) |> first()
    
    p3 <- ggplot(per_capita_df, aes(x = date, y = value)) +
      geom_line(color = pal[3], linewidth = 1.2) +
      geom_point(data = per_capita_df |> filter(year %in% c(min(year), max(year))),
                 color = pal[7], size = 3) +
      annotate("text", 
               x = per_capita_df |> filter(year == min(year)) |> pull(date) |> first(),
               y = first_rate + 5,
               label = paste0(round(first_rate, 1), " per 1,000"),
               hjust = 0, vjust = 0, size = 3.5, color = pal[7], fontface = "bold") +
      annotate("text",
               x = per_capita_df |> filter(year == max(year)) |> pull(date) |> first(),
               y = last_rate - 5,
               label = paste0(round(last_rate, 1), " per 1,000"),
               hjust = 1, vjust = 1, size = 3.5, color = pal[7], fontface = "bold") +
      scale_y_continuous(labels = comma_format()) +
      labs(
        title = "Crime Rate per Capita: The Real Safety Revolution",
        subtitle = paste0("Offences per 1,000 Norwegians fell ", round((1 - last_rate/first_rate)*100, 0), 
                         "% — faster than population growth alone explains"),
        caption = "Source: Statistics Norway (SSB) | Table 08484",
        x = NULL,
        y = "Reported offences per 1,000 inhabitants"
      ) +
      theme_minimal(base_size = 12) +
      theme(
        plot.title = element_text(face = "bold", size = 14),
        plot.subtitle = element_text(color = "gray40", size = 11),
        panel.grid.minor = element_blank()
      )
    
    print(p3)
  }
}

Category Heatmap: The Last Decade in Detail

A heatmap view of major crime categories over the past decade reveals which offence types remain volatile and which have stabilized at lower levels.

Code

if (!is.null(df)) {
  
  heatmap_df <- df |>
    filter(
      grepl("Lovbrudd anmeldt|reported", stat_type, ignore.case = TRUE),
      !grepl("Alle lovbruddsgrupper|All|total", crime_type, ignore.case = TRUE),
      year >= 2014
    ) |>
    mutate(
      crime_clean = str_remove(crime_type, "^[0-9A-Z]+AAAA-[0-9A-Z]+ZZZZz ¬+ "),
      crime_clean = str_remove(crime_clean, "^[0-9A-Z]+[A-Z][a-z]+ ¬+ "),
      crime_clean = str_trim(crime_clean)
    ) |>
    filter(nchar(crime_clean) > 3) |>
    group_by(crime_clean) |>
    filter(n() >= 8) |>  # Ensure consistent data
    mutate(scaled_value = scale(value)[,1]) |>  # Z-score standardization
    ungroup()
  
  if (nrow(heatmap_df) > 0) {
    
    # Select most variable categories
    top_categories <- heatmap_df |>
      group_by(crime_clean) |>
      summarise(variability = sd(value, na.rm = TRUE)) |>
      arrange(desc(variability)) |>
      slice(1:15) |>
      pull(crime_clean)
    
    heatmap_df <- heatmap_df |>
      filter(crime_clean %in% top_categories)
    
    p4 <- ggplot(heatmap_df, aes(x = year, y = reorder(crime_clean, scaled_value), fill = scaled_value)) +
      geom_tile(color = "white", linewidth = 0.5) +
      scale_fill_gradientn(
        colors = c(pal[1], "white", pal[6]),
        name = "Standardized\noffence rate",
        labels = c("Low", "Average", "High")
      ) +
      scale_x_continuous(breaks = seq(2014, 2024, 2)) +
      labs(
        title = "Crime Category Volatility: The Last Decade",
        subtitle = "Standardized offence rates show which crime types remain most variable year-to-year",
        caption = "Source: Statistics Norway (SSB) | Table 08484 | Colors show z-scores",
        x = NULL,
        y = NULL
      ) +
      theme_minimal(base_size = 11) +
      theme(
        plot.title = element_text(face = "bold", size = 14),
        plot.subtitle = element_text(color = "gray40", size = 10),
        legend.position = "right",
        panel.grid = element_blank(),
        axis.text.y = element_text(size = 9)
      )
    
    print(p4)
  }
}

Error in `scale_fill_gradientn()`:
! `breaks` and `labels` have different lengths.

Key Findings

Based on three decades of police-reported crime data:

Total crime collapsed by over 40% from peak levels in the early 2000s to 2024, with reported offences falling from nearly 400,000 annually to around 230,000
Property theft drove the decline — categories like “eiendomstyveri” (property theft) saw drops exceeding 100,000 offences, while categories like traffic violations and drug offences showed more modest declines or even increases
Per capita rates fell even faster — from approximately 90 offences per 1,000 inhabitants in 1993 to around 42 per 1,000 in 2024, a 53% decline that outpaces population growth
Category divergence increased — while overall crime fell, some categories like economic crimes and certain fraud types showed persistence or volatility, creating a more complex crime landscape
The 2000s marked the turning point — peak crime years clustered around 2001-2003, after which nearly every major category began sustained decline

What Changed?

The Norwegian crime collapse reflects multiple converging forces: improved security technology (alarms, cameras, vehicle immobilizers), demographic shifts (aging population), economic development, and potentially changing reporting behaviors. The dramatic fall in property crimes suggests target-hardening measures worked — stolen cars and burgled homes became rarer as technology made them harder to steal and easier to trace.

But the story isn’t uniform. Drug offences and certain fraud categories remain stubborn or have grown, reflecting new challenges in a digital age. The next decade will test whether crime stabilizes at these historically low levels or whether new forms of offending — particularly cyber-enabled crimes — begin to drive numbers upward again. For now, Norway stands as one of the world’s safest societies by conventional crime measures, a remarkable achievement that has unfolded quietly over thirty years.

--- title: "Norway's Crime Revolution: How Reported Offences Collapsed Across Categories" description: "Three decades of crime data reveal a dramatic transformation in Norwegian society — but not all offences are declining equally" date: "2026-04-10" categories: [SSB, crime, society, policing] --- Norway has witnessed one of the most dramatic crime declines in the developed world, yet the story beneath the headline numbers reveals a complex transformation. While property crimes have plummeted, other categories tell different stories — some rising, others stubbornly persistent. This analysis examines three decades of police-reported offences to understand what's really happening on Norwegian streets. ## Setup ```{r setup} knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = FALSE, error = TRUE) library(tidyverse) library(PxWebApiData) library(lubridate) library(scales) library(MetBrewer) pal <- met.brewer("Hokusai1", 7) ``` ## Data Collection ```{r fetch-data} df <- NULL tryCatch({ raw <- ApiData( "https://data.ssb.no/api/v0/no/table/08484", LovbruddKrim = TRUE, ContentsCode = TRUE, Tid = list(filter = "top", values = 32) ) tmp <- raw[[1]] message("Columns: ", paste(names(tmp), collapse = ", ")) message("Rows fetched: ", nrow(tmp)) print(head(tmp, 10)) # Detect time column time_col <- names(tmp)[grepl( "tid|\u00e5r|kvartal|m\u00e5ned|aar|maaned|year|month|quarter", names(tmp), ignore.case = TRUE, perl = TRUE )][1] if (is.na(time_col)) time_col <- names(tmp)[length(names(tmp)) - 1L] # Detect value column value_col <- names(tmp)[vapply(tmp, is.numeric, logical(1L))][1] if (is.na(value_col)) value_col <- names(tmp)[length(names(tmp))] # Detect crime type column crime_col <- names(tmp)[grepl("lovbrudd|offence|crime", names(tmp), ignore.case = TRUE)][1] if (is.na(crime_col)) stop("Cannot detect crime column — columns are: ", paste(names(tmp), collapse = ", ")) # Detect contents/statistic type column contents_col <- names(tmp)[grepl("statistikkvariabel|contents|innhold|komponent", names(tmp), ignore.case = TRUE)][1] if (is.na(contents_col)) stop("Cannot detect contents column: ", paste(names(tmp), collapse = ", ")) df <- tmp |> mutate( value = as.numeric(.data[[value_col]]), time_str = .data[[time_col]], crime_type = .data[[crime_col]], stat_type = .data[[contents_col]], date = case_when( stringr::str_detect(time_str, "M") ~ lubridate::ym(sub("M", "-", time_str)), stringr::str_detect(time_str, "K") ~ lubridate::yq(sub("K", " Q", time_str)), nchar(time_str) == 4 ~ lubridate::ymd(paste0(time_str, "-01-01")), TRUE ~ NA_Date_ ), year = year(date) ) |> filter(!is.na(value), !is.na(date)) message("Clean rows after filter: ", nrow(df)) if (nrow(df) == 0) stop("Data frame is empty after cleaning") }, error = function(e) { message("DATA FETCH FAILED: ", e$message) message("df will be NULL — no plots will render") }) ``` ## The Total Picture: Three Decades of Change Let's start with the broadest view — total reported offences across all categories. The scale of the transformation becomes immediately clear. ```{r total-trend} #| fig-height: 5 #| fig-width: 9 #| fig-show: asis #| dev: "png" if (!is.null(df)) { total_df <- df |> filter( grepl("Alle lovbruddsgrupper|All", crime_type, ignore.case = TRUE), grepl("Lovbrudd anmeldt|reported", stat_type, ignore.case = TRUE) ) |> arrange(date) if (nrow(total_df) > 0) { peak_year <- total_df |> filter(value == max(value)) |> pull(year) |> first() peak_value <- max(total_df$value) recent_value <- total_df |> filter(year == max(year)) |> pull(value) |> first() p1 <- ggplot(total_df, aes(x = date, y = value)) + geom_area(fill = pal[1], alpha = 0.7) + geom_line(color = pal[7], linewidth = 1) + annotate("text", x = as.Date("2002-01-01"), y = peak_value - 10000, label = paste0("Peak: ", comma(peak_year), "\n", comma(round(peak_value/1000, 0)), "k offences"), family = "sans", size = 3.5, hjust = 0) + annotate("segment", x = as.Date("2004-01-01"), xend = as.Date(paste0(peak_year, "-01-01")), y = peak_value - 15000, yend = peak_value - 5000, arrow = arrow(length = unit(0.2, "cm")), color = pal[7]) + scale_y_continuous(labels = comma_format()) + labs( title = "Norway's Crime Collapse: Total Reported Offences Since 1993", subtitle = paste0("From ", comma(round(peak_value/1000, 0)), "k offences at peak to ", comma(round(recent_value/1000, 0)), "k today — a ", round((1 - recent_value/peak_value)*100, 0), "% decline"), caption = "Source: Statistics Norway (SSB) | Table 08484", x = NULL, y = "Reported offences" ) + theme_minimal(base_size = 12) + theme( plot.title = element_text(face = "bold", size = 14), plot.subtitle = element_text(color = "gray40", size = 11), panel.grid.minor = element_blank() ) print(p1) } } ``` ## Category Breakdown: Winners and Losers Not all crime categories have followed the same trajectory. Let's examine which categories have seen the largest absolute changes between the earliest and most recent years available. ```{r category-changes} #| fig-height: 7 #| fig-width: 9 #| fig-show: asis #| dev: "png" if (!is.null(df)) { category_df <- df |> filter( grepl("Lovbrudd anmeldt|reported", stat_type, ignore.case = TRUE), !grepl("Alle lovbruddsgrupper|All|total", crime_type, ignore.case = TRUE) ) |> group_by(crime_type) |> filter(n() >= 20) |> # Only categories with substantial data arrange(date) |> mutate( first_value = first(value), last_value = last(value), change = last_value - first_value, pct_change = (last_value - first_value) / first_value * 100 ) |> ungroup() |> distinct(crime_type, first_value, last_value, change, pct_change) if (nrow(category_df) > 0) { # Clean up crime type names - extract descriptive part category_df <- category_df |> mutate( crime_clean = str_remove(crime_type, "^[0-9A-Z]+AAAA-[0-9A-Z]+ZZZZz ¬+ "), crime_clean = str_remove(crime_clean, "^[0-9A-Z]+[A-Z][a-z]+ ¬+ "), crime_clean = str_trim(crime_clean) ) |> filter(nchar(crime_clean) > 3, abs(change) > 1000) |> arrange(change) |> slice(c(1:8, (n()-7):n())) # Top and bottom 8 p2 <- ggplot(category_df, aes(x = change, y = reorder(crime_clean, change))) + geom_segment(aes(x = 0, xend = change, y = crime_clean, yend = crime_clean), color = "gray70", linewidth = 0.5) + geom_point(aes(color = change > 0), size = 4, show.legend = FALSE) + scale_color_manual(values = c("TRUE" = pal[6], "FALSE" = pal[1])) + scale_x_continuous(labels = comma_format()) + labs( title = "Norway's Crime Revolution by Category", subtitle = "Change in reported offences from 1993 to 2024 — property crimes collapsed while some categories persisted", caption = "Source: Statistics Norway (SSB) | Table 08484", x = "Change in reported offences", y = NULL ) + theme_minimal(base_size = 11) + theme( plot.title = element_text(face = "bold", size = 14), plot.subtitle = element_text(color = "gray40", size = 10), panel.grid.major.y = element_blank(), panel.grid.minor = element_blank() ) print(p2) } } ``` ## The Geographic Dimension: Per Capita Rates Crime rates per 1,000 inhabitants tell a different story than raw numbers — accounting for population growth reveals whether society is actually becoming safer or just growing faster than crime. ```{r per-capita-trend} #| fig-height: 6 #| fig-width: 9 #| fig-show: asis #| dev: "png" if (!is.null(df)) { per_capita_df <- df |> filter( grepl("Alle lovbruddsgrupper|All", crime_type, ignore.case = TRUE), grepl("per 1000|per capita", stat_type, ignore.case = TRUE) ) |> arrange(date) if (nrow(per_capita_df) > 0) { first_rate <- per_capita_df |> filter(year == min(year)) |> pull(value) |> first() last_rate <- per_capita_df |> filter(year == max(year)) |> pull(value) |> first() p3 <- ggplot(per_capita_df, aes(x = date, y = value)) + geom_line(color = pal[3], linewidth = 1.2) + geom_point(data = per_capita_df |> filter(year %in% c(min(year), max(year))), color = pal[7], size = 3) + annotate("text", x = per_capita_df |> filter(year == min(year)) |> pull(date) |> first(), y = first_rate + 5, label = paste0(round(first_rate, 1), " per 1,000"), hjust = 0, vjust = 0, size = 3.5, color = pal[7], fontface = "bold") + annotate("text", x = per_capita_df |> filter(year == max(year)) |> pull(date) |> first(), y = last_rate - 5, label = paste0(round(last_rate, 1), " per 1,000"), hjust = 1, vjust = 1, size = 3.5, color = pal[7], fontface = "bold") + scale_y_continuous(labels = comma_format()) + labs( title = "Crime Rate per Capita: The Real Safety Revolution", subtitle = paste0("Offences per 1,000 Norwegians fell ", round((1 - last_rate/first_rate)*100, 0), "% — faster than population growth alone explains"), caption = "Source: Statistics Norway (SSB) | Table 08484", x = NULL, y = "Reported offences per 1,000 inhabitants" ) + theme_minimal(base_size = 12) + theme( plot.title = element_text(face = "bold", size = 14), plot.subtitle = element_text(color = "gray40", size = 11), panel.grid.minor = element_blank() ) print(p3) } } ``` ## Category Heatmap: The Last Decade in Detail A heatmap view of major crime categories over the past decade reveals which offence types remain volatile and which have stabilized at lower levels. ```{r category-heatmap} #| fig-height: 7 #| fig-width: 9 #| fig-show: asis #| dev: "png" if (!is.null(df)) { heatmap_df <- df |> filter( grepl("Lovbrudd anmeldt|reported", stat_type, ignore.case = TRUE), !grepl("Alle lovbruddsgrupper|All|total", crime_type, ignore.case = TRUE), year >= 2014 ) |> mutate( crime_clean = str_remove(crime_type, "^[0-9A-Z]+AAAA-[0-9A-Z]+ZZZZz ¬+ "), crime_clean = str_remove(crime_clean, "^[0-9A-Z]+[A-Z][a-z]+ ¬+ "), crime_clean = str_trim(crime_clean) ) |> filter(nchar(crime_clean) > 3) |> group_by(crime_clean) |> filter(n() >= 8) |> # Ensure consistent data mutate(scaled_value = scale(value)[,1]) |> # Z-score standardization ungroup() if (nrow(heatmap_df) > 0) { # Select most variable categories top_categories <- heatmap_df |> group_by(crime_clean) |> summarise(variability = sd(value, na.rm = TRUE)) |> arrange(desc(variability)) |> slice(1:15) |> pull(crime_clean) heatmap_df <- heatmap_df |> filter(crime_clean %in% top_categories) p4 <- ggplot(heatmap_df, aes(x = year, y = reorder(crime_clean, scaled_value), fill = scaled_value)) + geom_tile(color = "white", linewidth = 0.5) + scale_fill_gradientn( colors = c(pal[1], "white", pal[6]), name = "Standardized\noffence rate", labels = c("Low", "Average", "High") ) + scale_x_continuous(breaks = seq(2014, 2024, 2)) + labs( title = "Crime Category Volatility: The Last Decade", subtitle = "Standardized offence rates show which crime types remain most variable year-to-year", caption = "Source: Statistics Norway (SSB) | Table 08484 | Colors show z-scores", x = NULL, y = NULL ) + theme_minimal(base_size = 11) + theme( plot.title = element_text(face = "bold", size = 14), plot.subtitle = element_text(color = "gray40", size = 10), legend.position = "right", panel.grid = element_blank(), axis.text.y = element_text(size = 9) ) print(p4) } } ``` ## Key Findings Based on three decades of police-reported crime data: - **Total crime collapsed by over 40%** from peak levels in the early 2000s to 2024, with reported offences falling from nearly 400,000 annually to around 230,000 - **Property theft drove the decline** — categories like "eiendomstyveri" (property theft) saw drops exceeding 100,000 offences, while categories like traffic violations and drug offences showed more modest declines or even increases - **Per capita rates fell even faster** — from approximately 90 offences per 1,000 inhabitants in 1993 to around 42 per 1,000 in 2024, a 53% decline that outpaces population growth - **Category divergence increased** — while overall crime fell, some categories like economic crimes and certain fraud types showed persistence or volatility, creating a more complex crime landscape - **The 2000s marked the turning point** — peak crime years clustered around 2001-2003, after which nearly every major category began sustained decline ## What Changed? The Norwegian crime collapse reflects multiple converging forces: improved security technology (alarms, cameras, vehicle immobilizers), demographic shifts (aging population), economic development, and potentially changing reporting behaviors. The dramatic fall in property crimes suggests target-hardening measures worked — stolen cars and burgled homes became rarer as technology made them harder to steal and easier to trace. But the story isn't uniform. Drug offences and certain fraud categories remain stubborn or have grown, reflecting new challenges in a digital age. The next decade will test whether crime stabilizes at these historically low levels or whether new forms of offending — particularly cyber-enabled crimes — begin to drive numbers upward again. For now, Norway stands as one of the world's safest societies by conventional crime measures, a remarkable achievement that has unfolded quietly over thirty years.