SPS Management Information Technical Manual

Daily snapshots are used to produce a number of data products, detailed below.

Daily snapshots

Daily snapshots are typically extracted from PR2 just after midnight each day. Records included are those whose “In Custody” variable has the value TRUE. There are known shortcomings to this variable, which is not automatically updated when a prisoner is admitted or assigned a cell[i]. The analytical team receives a daily report of these infrequent discrepancies, but they are corrected by system administrators after the extract of the snapshot data.

Compiled daily snapshots (“spsMI”)

Daily snapshots are collated for comparison between days and for analysis.

Comparing prisoner manifests in each snapshot allows us to chart changes in the prison population and its characteristics. We can also identify populations flows, comparing snapshots for presence/absence of individuals, and status changes by comparing legal status between snapshots.

Because of the 24-hourly capture period, this collection will not capture individuals who enter and leave in the space of a single day.

Longitudinalised snapshots (“longSnap”)

Daily snapshots are compiled in our longSnap dataset, which is used to determine short-term trends from the management information[ii]. This allows us to identify periods of continuous occupancy and changes in legal status between remand and sentenced over the span of the dataset.

In contrast to the cellWise data[iii], retrograde transitions (from sentenced to remand) can be identified using longSnap provided at least one night is spent in custody with that new status.

Identifying continuous presence from the pre-pandemic period requires the addition of cellWise data prior to 28th March 2020.

As with the spsMI flow data, this dataset will miss individuals who enter and leave custody in the space of a day. It will also miss legal status changes (R -> S or S -> R) that do not persist overnight.

Population Modelling Data

Monthly compilations of daily snapshots are used to track flows in the prison population by Operational Researchers in their production of the short-term prison population model.

In the microsimulation model, prison population data is used to set the initial size of the different prison sub-populations, such as sentenced and remand. This dataset also provides the arrival dates and overall sentence lengths used in the model.[iv]