Visualisation of 19th Century Women Prisoner Data

In this work-in-progress project we visualise a sample of transcribed data from a large spreadsheet. The data is from prison records of women incarcerated in Victoria between 1885 and 1920. The backgrounds and conviction histories of the women vary. Some women were imprisoned only once, but others returned to gaol again and again, with the most prolific offender amassing 188 convictions on her prison record. The vast bulk of women were imprisoned for petty offences such as vagrancy and public drunkenness, but others were convicted of serious property or violent crimes. Women from all walks of life ended up in Victoria's prisons, but Irish women and women employed as domestic servants are notably over-represented in the data. Analysing patterns in the data can enable greater understanding of the different factors that contributed to women's imprisonment across time.

This data is represented visually using the popular Parallel Coordinates method. The data has been provided by Dr Alana Jayne Piper from the Australian Centre for Public History here at UTS. The full circumference of the Data Arena's display has been used to show the original spreadsheet and roughly 30 dimensions / columns represented as axes around the room.

Spreadsheet Data viewed in Parallel Coordinates

Parallel Coordinates? Here's how the works. Each column in the spreadsheet is represented visually as a vertical axis. The column's range of values appear as tick-marks on the axis. If the spreadsheet column contained values, like the year of birth, then that column has the various years marked out, from the earliest year at the bottom to the most recent year at the top. If instead the spreadsheet column contained some categorical information, such as Religion, then these are laid out top-to-bottom in alphabetical order. All the axes are vertical - hence parallel.

Each line is a Convict

Take another look at the graph. The coloured horizontal lines represent each row of the spreadsheet - indeed each convict prisoner. One coloured line is one prisoner. Like a piece of wool wired up to nails on each column, the lines are drawn to cross each vertical axis at the value appropriate - from the value in the spreadsheet. Thus a 1820 convict who was Roman Catholic, would be represented by a line passing through those tick marks on the Date & Religion columns. Doing this, one line for each convict then shows where some values are common, and where some are outliers.

Filter, explore ...

Further, we can drill down on this data. By selection subregions of a particular column, say only 1820 to 1830 in the year axis, we see particular lines highlighted - the lines which pass through this date range appear brighter. This allows the user to see what other axes (convict details) may correlate or not correlate which that selection. Further regions on other axes can refine the search, say by then selecting a particular religion, or a sub-region on any other axis.

For more information on this topic, take a look at Dr Alana Jayne Piper's website Criminal Characters