Comprehensive Guide to Displaying and Visualising Data in Criminology and Statistics

Foundations of Data Visualisation

  • Core Purpose of Visualising Data:

    • Generates insight: Translates complex numerical values into an interpretable format that facilitates rapid pattern recognition and cognitive processing.

    • Uncovers underlying structure: Exposes organizational tendencies, latent shapes, groupings, and symmetries within datasets that remain obscured in raw rows and columns.

    • Identifies important variables: Visual exploration highlights which parameters show meaningful variation, strong associations, or substantial influence on outcomes of interest.

    • Detects outliers and anomalies: Visually isolates erroneous records, extreme events, data-entry errors, and irregular distributions requiring imputation or verification.

    • Helps make sense of patterns: Converts abstract empirical observations into discernable trajectories, trends, seasonal shifts, and cluster configurations.

Frequency Distributions and Tabular Summaries

  • Conceptual Definitions:

    • Frequency Distribution Table: A structured summary table displaying the record count (number of individuals or events) distributed across distinct mutually exclusive categories or intervals.

    • Absolute Frequency (ff): The raw count of observations occurring within a designated category or bin interval.

    • Relative Frequency (rfrf): The proportion or percentage of the total sample size (NN) that falls within a given category:         Relative Frequency=(fN)×100\text{Relative Frequency} = \left( \frac{f}{N} \right) \times 100

    • Cumulative Frequency (cfcf): The successive running total of absolute frequencies accumulated from the initial category down through the current category:         cfi=∑j=1ifjcf_i = \sum_{j=1}^{i} f_j

    • Cumulative Relative Frequency (crfcrf): The successive running sum of relative frequencies expressed as an accumulated percentage of the total distribution:         crfi=∑j=1irfj=(cfiN)×100crf_i = \sum_{j=1}^{i} rf_j = \left( \frac{cf_i}{N} \right) \times 100

  • Case Demonstration: Fortitude Valley Police Offence Data:

    • Dataset Characteristics: A sample of raw recorded police incidents (N=611N = 611 records) occurring between 24 February 202224\text{ February } 2022 and 23 March 202223\text{ March } 2022 in Fortitude Valley.

    • Analytical Tool: Tabulation executed via Pivot Tables in spreadsheet software (such as Microsoft Excel) to aggregate individual offence occurrences into grouped frequencies.

    • Complete Offence Distribution Metrics (N=611N = 611):

      • Good Order Offences: Absolute Frequency = 168168; Relative Frequency = 27.50%27.50\%; Cumulative Frequency = 334334; Cumulative Relative Frequency = 54.66%54.66\%

      • Other Theft (excl. Unlawful Entry): Absolute Frequency = 142142; Relative Frequency = 23.24%23.24\%; Cumulative Frequency = 524524; Cumulative Relative Frequency = 85.76%85.76\%

      • Drug Offences: Absolute Frequency = 117117; Relative Frequency = 19.15%19.15\%; Cumulative Frequency = 159159; Cumulative Relative Frequency = 26.02%26.02\%

      • Assault: Absolute Frequency = 4242; Relative Frequency = 6.87%6.87\%; Cumulative Frequency = 4242; Cumulative Relative Frequency = 6.87%6.87\%

      • Traffic & Related Offences: Absolute Frequency = 3232; Relative Frequency = 5.24%5.24\%; Cumulative Frequency = 559559; Cumulative Relative Frequency = 91.49%91.49\%

      • Other Property Damage: Absolute Frequency = 3030; Relative Frequency = 4.91%4.91\%; Cumulative Frequency = 382382; Cumulative Relative Frequency = 62.52%62.52\%

      • Unlawful Entry: Absolute Frequency = 2323; Relative Frequency = 3.76%3.76\%; Cumulative Frequency = 587587; Cumulative Relative Frequency = 96.07%96.07\%

      • Weapons Act Offences: Absolute Frequency = 1313; Relative Frequency = 2.13%2.13\%; Cumulative Frequency = 611611; Cumulative Relative Frequency = 100.00%100.00\%

      • Unlawful Use of Motor Vehicle: Absolute Frequency = 1111; Relative Frequency = 1.80%1.80\%; Cumulative Frequency = 598598; Cumulative Relative Frequency = 97.87%97.87\%

      • Fraud: Absolute Frequency = 77; Relative Frequency = 1.15%1.15\%; Cumulative Frequency = 166166; Cumulative Relative Frequency = 27.17%27.17\%

      • Handling Stolen Goods: Absolute Frequency = 66; Relative Frequency = 0.98%0.98\%; Cumulative Frequency = 340340; Cumulative Relative Frequency = 55.65%55.65\%

      • Other Offences Against the Person: Absolute Frequency = 66; Relative Frequency = 0.98%0.98\%; Cumulative Frequency = 352352; Cumulative Relative Frequency = 57.61%57.61\%

      • Trespassing & Vagrancy: Absolute Frequency = 55; Relative Frequency = 0.82%0.82\%; Cumulative Frequency = 564564; Cumulative Relative Frequency = 92.31%92.31\%

      • Liquor (excl. Drunkenness): Absolute Frequency = 44; Relative Frequency = 0.65%0.65\%; Cumulative Frequency = 344344; Cumulative Relative Frequency = 56.30%56.30\%

      • Robbery: Absolute Frequency = 33; Relative Frequency = 0.49%0.49\%; Cumulative Frequency = 527527; Cumulative Relative Frequency = 86.25%86.25\%

      • Miscellaneous Offences: Absolute Frequency = 22; Relative Frequency = 0.33%0.33\%; Cumulative Frequency = 346346; Cumulative Relative Frequency = 56.63%56.63\%

      • Grand Total: Absolute Frequency = 611611; Relative Frequency = 100.00%100.00\%

Comprehensive frequency distribution table for offence types

Categorical Data Visualisation: Bar Charts

  • Structural Mechanics of Bar Charts:

    • Designed specifically for nominal and ordinal categorical data.

    • Each distinct category is assigned an individual, discrete vertical or horizontal bar.

    • A continuous axis represents quantitative magnitude, where the height (or length) of each bar corresponds directly to its absolute count or relative percentage frequency.

    • Deliberate spaces exist between adjacent bars to visually communicate that categories are qualitative and discrete rather than continuous.

  • Presentation and Ordering Strategies:

    • Default Alphabetical Ordering: Sorting categories alphabetically (e.g., Assault, Drug Offences, Fraud…) produces erratic variations in bar height across the horizontal axis, making global ranking and visual estimation cumbersome.

    • Frequency-Ordered (Pareto) Layout: Reordering categories from highest frequency down to lowest frequency (e.g., Good Order Offences [168168], Other Theft [142142], Drug Offences [117117] down to Miscellaneous Offences [22]) provides immediate clarity regarding modal offence types and relative shares.

Alphabetical bar chart of crime counts by offence typeReordered bar chart showing offence types sorted from highest to lowest frequency

Continuous Data Visualisation: Histograms

  • Structural Mechanics of Histograms:

    • Utilised exclusively for continuous interval and ratio numeric variables.

    • Constructed using adjacent rectangles with zero spacing between boundaries, reflecting the continuous, unbroken nature of the measurement scale.

    • The total range of continuous data is divided into contiguous, non-overlapping intervals termed bins or class intervals.

  • Diagnostic Capabilities:

    • Data Integrity and Error Detection: Quickly reveals recording errors, invalid negative codes (e.g., an age value coded as −2-2 denoting missing data or a system error), and out-of-range figures.

    • Distributional Profile and Central Tendency: Displays skewness (positive right skew, negative left skew), central clustering, kurtosis, and the distribution's mode (the highest vertical bar).

  • Binning Considerations (World Values Survey 2018 Case Study):

    • Dataset Scope: Continuous age variable across 17951795 survey respondents (N=1795N = 1795).

    • Data Range: Minimum recorded valid age is 1717; maximum recorded age is 9898.

    • Necessity of Binning: Plotting all individual single-year data points would generate 8181 discrete values, creating visual noise with negligible informational value.

    • Class Interval Construction: Data are aggregated into grouped bins across age ranges: 0–200\text{--}20, 21–3021\text{--}30, 31–4031\text{--}40, 41–5041\text{--}50, 51–6051\text{--}60, 61–7061\text{--}70, 71–8071\text{--}80, 81–9081\text{--}90, and 91–10091\text{--}100.

    • Distributional Outcome: The respondent age histogram reveals a peak clustering in the middle-to-older adult brackets, with the modal frequency occurring in the 51–6051\text{--}60 age cohort (approaching 400400 respondents), followed by the 41–5041\text{--}50 cohort.

Histogram of respondent age distribution
  • Skewness and Outlier Visualisation (Brisbane Violent Crime Study):

    • Comparative longitudinal analysis across Brisbane neighbourhoods for years 20012001, 20062006, and 20112011.

    • Shape: Exhibits extreme positive (right-hand) skewness. The overwhelming majority of neighbourhoods record between 00 and 2020 violent crimes.

    • Outlier Trajectory: Extended right-hand tails stretch toward counts of 100100, 150150, and 200200 incidents, visually exposing a small subset of severe crime hot spots across all three evaluated census years.

Distribution of violent crime counts across Brisbane neighbourhoods in 2001, 2006, and 2011

Longitudinal Data Visualisation: Line Charts

  • Characteristics and Usage:

    • Continuous line charts connect individual quantitative measurements across an ordered sequence—most commonly successive temporal increments (e.g., weeks, months, years).

    • Ideal for displaying trajectories, rate of change, cyclic trends, convergence, and divergence between distinct population subgroups.

  • Empirical Demonstration: Queensland Imprisonment Rates (2011–20212011\text{--}2021):

    • Metric: Imprisonment rates per 100,000100,000 population segmented by sex across Queensland correctional facilities.

    • Annual Historical Rates (2011–20212011\text{--}2021):

      • 20112011: Male = 302.1302.1; Female = 24.024.0

      • 20122012: Male = 293.7293.7; Female = 26.726.7

      • 20132013: Male = 311.4311.4; Female = 29.529.5

      • 20142014: Male = 353.5353.5; Female = 36.236.2

      • 20152015: Male = 361.9361.9; Female = 37.737.7

      • 20162016: Male = 382.0382.0; Female = 35.835.8

      • 20172017: Male = 415.1415.1; Female = 36.036.0

      • 20182018: Male = 419.8419.8; Female = 42.242.2

      • 20192019: Male = 415.0415.0; Female = 42.842.8

      • 20202020: Male = 406.6406.6; Female = 37.737.7

      • 20212021: Male = 460.7460.7; Female = 45.245.2

    • Comparative Insights:

      • Scale Disparity: Male imprisonment rates reside on a substantially higher baseline (293.7–460.7293.7\text{--}460.7 per 100,000100,000) compared to female imprisonment rates (24.0–45.224.0\text{--}45.2 per 100,000100,000).

      • Temporal Progression: Both demographics show a marked net upward trajectory over the decade, with male incarceration climbing by over 150150 points per 100,000100,000 and female incarceration nearly doubling.

Chart Selection Framework: The Chart Chooser Taxonomy

  • Decision Foundation:

    • Developed by Dr. Andrew V. Abela (20202020), the taxonomy provides a decision matrix to identify appropriate visualisation tools based on communicative intent: "What would you like to show?"

    • The framework segments visualisations into four distinct communicative categories: Comparison, Distribution, Composition, and Relationship.

  • Category 1: Comparison:

    • Static Context (No Time Dimension):

      • One Variable: Single Bar Chart.

      • Two Variables: Variable Width Column Chart.

      • Three or More Variables: Multiple Bar Charts (few items) or Table / Table with Embedded Charts (many items).

    • Changing Over Time Context:

      • Many Periods: Circular Area Chart (for cyclical data) or Line Chart (for non-cyclical data).

      • Few Periods: Column Chart (single or few items) or Line Chart (many items).

  • Category 2: Distribution:

    • Single Variable: Column Histogram (few items) or Line Histogram (many items / continuous profiles).

    • Two Variables: Scatter Chart.

    • Three Variables: 3D Area Chart.

  • Category 3: Composition:

    • Static Composition:

      • Simple Share of Total: Pie Chart.

      • Accumulation or Subtraction to Total: Waterfall Chart.

      • Components of Components: Stacked 100%100\% Column Chart with Subcomponents.

    • Changing Over Time:

      • Few Periods: Stacked 100%100\% Column Chart (when only relative differences matter) or Stacked Column Chart (when relative and absolute differences both matter).

      • Many Periods: Stacked 100%100\% Area Chart (when only relative differences matter) or Stacked Area Chart (when relative and absolute differences both matter).

  • Category 4: Relationship:

    • Two Continuous Variables: Scatter Chart.

    • Three Continuous Variables: Bubble Chart (where the third variable dictates bubble diameter).

Chart Chooser decision taxonomy diagram by Andrew V. Abela