Abstract
Introduction
Organisations are increasingly being driven by large volumes of data and information, and the technologies that allow them to turn these into useful knowledge, and organisational wisdom. Across all sectors, the transformative potential of data analytics is being used to better understand and improve organisational decision-making (Gandomi and Haider, 2015). One such area that has made use of this access to systems that can organise and help manage large amounts of data is law enforcement, where police forces are starting to harness data analytics to allow them to better understand their communities, comprehend the patterns of crime and criminality, improve operational efficiency, and ultimately, make communities safer.
Crime is a complex phenomenon, driven by a range of factors. Police forces in the UK work with a large range of other organisations, such as health services, housing, fire and rescue services, charities, children’s homes and other law enforcement agencies. These partnerships and collaborations are themselves complex and rely on good sharing of information (Zaghloul and Partridge, 2022). Police forces collect a wide range of data generated by themselves and these other organisations. Examples might include direct information on crimes and criminals generated by public calls to emergency services and police investigations. It also includes data and intelligence from other organisations such as data on hospital admissions for knife injuries, what types of drugs have been seized by customs and borders agencies, information on children at risk, intelligence on serious criminal activity, and intelligence on low level anti-social behaviour. Police forces also gather intelligence on their local communities – where do people feel unsafe, what criminal issues are being faced by each community, as well as publicly available information (e.g. social media posts) to help identify criminals and criminal activity (Williams et al., 2018; Cano-Marin et al., 2023).
More recent technologies such as facial recognition, video surveillance and artificial intelligence (AI) have opened the potential for even more data to be gathered, but could also create an overload of data that cannot be assessed and turned into useful knowledge in an effective manner. This data overload poses a significant challenge, but also presents an opportunity to extract valuable insights and patterns that can be used to inform decision-making and allocate resources more effectively.
This case study looks at data analytics within a police force, examining how they recognised the need to make better use of the data and information they held, and how they started to professionalise their data analytical capability. This real-world example of the decisions and issues faced by a police force in the UK when looking at data analytics will explore the impact of data analytics on crime prevention, resource allocation, and community engagement, and the associated issues faced by a law enforcement body around culture, ethics and technology when implementing such an approach. The case expects students to (a) understand data analytics adoption in law enforcement, (b) consider the challenges associated with data analytics adoption and (c) explore and be able to identify the similarities and differences between law enforcement and other industries.
Data analytics adoption in public services
Data analytics in policing could involve the use of sophisticated algorithms, statistical models, and machine learning techniques to extract, analyse, and interpret data, and develop this into knowledge that can drive action. By integrating a range of different data source and using advanced analytical tools, police forces could improve their understanding of risk, crime patterns, key hotspots and individuals to help predict potential incidents, and develop targeted intervention strategies.
Furthermore, data analytics could be used to optimise how scarce resources such as police officers are used. A better understanding of historical calls from the public, and crime data, mapped to geographical areas could be used to more effectively allocate police officers and other staff both geographically and temporally across days of the week and hours of the day. Efforts could be targeted to high-crime areas, high-risk crimes and key times of the day to improve the safety of communities and help reduce crime.
Done well, data analytics can improve the trust of communities. Trust in the police is an important part of an effective crime reduction approach. If the public have trust in the police they are more likely to report crimes, give evidence and engage with crime reduction advice. Using data-driven insights, police forces could proactively address community concerns and tailor the policing approach to specific demographic groups. Data-driven transparency initiatives like crime maps and statistical reports can be used to improve accountability, empower citizens and promote collaborative problem-solving between law enforcement agencies and the communities they serve.
However, there are also concerns about privacy, bias, and data security from data analytics adoption, and the use of AI, and algorithms. This case study will also consider ethical and legal dimensions of data analytics, and the importance of data governance, privacy safeguards and the need for ongoing public dialogue.
The context of policing in the UK
Police forces in the UK deal with a range of issues, from serious crimes such as murder, rape and burglary, through to anti-social behaviour such as harassment and noisy neighbours. Calls from the public come in on either the emergency 999 number, or the non-emergency 101 number, and are assessed by trained call takers who take details and record these on computer systems. Calls are then graded for an appropriate response that ranges from emergency (police aim to attend within a few minutes), to telephone resolution (the issue can be resolved by a telephone call). Many calls also do not require any police action and are passed to other agencies to deal with. Appropriate resources are then assigned to each call, depending on the grade and what resources are available in the area at the time.
Most police forces measure their attendance times for each grade of call. All forces are assessed on how well they identify and classify diverse types of crime, and how well they investigate and keep victims informed of progress. Statistics are also kept on the rate of different crime types, as the UK model of policing is focussed on crime prevention as well as detecting crimes that have occurred.
Neighbourhood policing teams tend to deal with preventing crime. The force uses crime trends (which types of crime seem to be increasing), hotspot locations (are there particular streets, areas or villages which have more crime than others) and known individuals (is a prolific burglar about to be released from prison for example). The police use their data and work jointly with partners and their data to understand these trends and likely causes of crime, to agree action plans to prevent crimes occurring.
In more serious cases, the police have detailed multi-agency meetings with other agencies to manage people who are particularly vulnerable to being victims of crime, or to manage dangerous offenders. Police data is again used in conjunction with data from other agencies such as housing, drug treatment charities, health and education to identify the best ways to keep the public safe and reduce crime.
All police data and data shared by partners is managed on a range of IT systems. Some are dedicated systems to manage calls from the public, to record the details of investigations or keep track of intelligence. Others are more ad-hoc, spreadsheet-based systems developed by people in the force to track and manage key issues. Police forces, like other agencies in the UK, are bound by legislation on how they use and retain data. Data must be for a policing purpose and must be deleted once no longer needed (College of Policing, 2023).
Increasingly police forces are making use of automated approaches that can search multiple data sources and connect seemingly disparate individuals or crimes. They are also using new technologies such as facial recognition, gait recognition (identifying individuals from the unique way they walk) and automated analysis of text, images and videos seized from suspects or posted on the Internet that can flag up potentially illegal content. Many of these tactics are controversial – there have been protests about police use of facial identification technology for example.
Summary of the case
A medium-sized UK police force was struggling with the volume of data that was generated from their key systems, and from partners. This data covered a range of areas, from crime reports and intelligence submitted by their own workforce, to data on criminality and risk provided from partners such as hospitals, social services, children’s services and more. The data was often in poor condition, with errors in how it was captured, duplications in submissions, incompatible formats and sheer volume meaning that gaining insights from it was time consuming and required significant manual collation and presentation of the data.
The police force was considering how this could be improved, using a dedicated data analytics team and better use of technological solutions. The costs of this approach however needed to be weighed against overall budget reductions 1 and increasingly complex crime investigations, which meant that any investment in data analytics needed to be justifiable. This case study looks at the proposals for improved data analytics in a UK police force, and covers issues of legality, ethics, technology as well as the analysis of the data.
The case: description
The new Head of Performance was concerned. She had just taken up the role in a medium-sized police force, but thought that the connection between the performance information that the police force leadership team used to make decisions and the data that the organisation gathered was weak. She raised her concerns with her Head of Department: We collect a lot of data, but we don’t really make use of most of it. My team produce two types of reports – the monthly, quarterly and annual performance reports that look at trends in volumes of crime, different crime types and measures such as prosecutions, and ad hoc reports where neighbourhood officers have a concern about a particular crime or issue and want some data to support their actions.
She explained: The first type of report is very high level and generic – it doesn’t really assist senior officers to make decisions, and is largely based on historical data, and not predictions of what might happen in the near future. For some crime types that don’t occur very often, one or two instances of a crime can also skew the figures. This can lead to resources being allocated inefficiently. The ad hoc reports are much better, as we can work with the local neighbourhood teams to understand their concerns, and can use the data we have from ourselves and key partners to identify historical trends and predict these into the future. For example, where hospitals are seeing an increase in people with knife injuries treated in emergency rooms, we can match this to reports of anti-social behaviour and gang activity in areas to indicate that there may be a gang related problem starting to emerge, and use specialist teams to address this before it becomes a major issue.
However, The problems with these ad hoc requests are twofold. Firstly they are resource intensive to produce. I have a small team with no dedicated people for this work, and they have to manually look through the various systems that hold the data, clean and format it so it can be analysed and then analyse and present it using the only tool we have, Microsoft Excel spreadsheets. Secondly, not all people in the force ask for these ad hoc requests – or if they do they already have an answer in their head and just want the data to support their views, even where it doesn’t.
As a result of these concerns, her Head of Department asked her to develop some proposals for how the force could make better use of the data that they have, and to justify why this was something worth spending their limited budget on. If he is convinced by the report, he will take it to the Force Resourcing Board who decide upon investments, commission projects and allocate resources.
Overview of the current approach to analysing data
The Head of Performance has only just taken up her post, and has a small team of 6 people who use spreadsheets to produce management information and performance reports. These are produced monthly, quarterly and annually, and are used for both decision-making and performance management internally, and (in summary form) to publish externally. These are seen as critical, and most of the team’s time is spent on extracting data, analysis and the presentation of these reports.
On a fairly regular basis, some police officers come to the team and ask them to look at emerging or concerning issues in their area. These requests can cover any type of issue, from gangs in neighbourhoods, to the prevalence of young people going missing from care homes, to thefts of garden furniture to reports of fraud – in effect covering the whole range of police activity.
She has noted that her team does a good job in producing these reports using a range of police, partner and public data to look at historical trends and help predict potential future issues. However, the team has no dedicated resources to support this work, and has to manually access lots of different systems to extract the data, and then clean this data to a standardised format, before importing it to a spreadsheet and conducting the analysis using basic office tools. This is both time consuming and prone to technology issues as the computers they use cannot always manage the volume of data and regularly crash.
She is aware of other police forces that are working with some local universities to train their analytical staff on the use of more appropriate data analytical methods and tools. Her force is a member of this group, but because they don’t have a dedicated analytical team, have not really participated in this group. The group has trained staff in other forces to use statistical tools such as R and SPSS and to use data analytical methods to automate and improve their reporting. She seeks the views of her team on the matter. The reports that I run on Excel can take hours to complete. I have automated the import and some of the analysis, but the spreadsheets we use can then take ages to calculate and refresh the reports. I usually set these running last thing at night, but more and more I am finding that something went wrong, and the system crashed meaning I have to redo much of the work. (Team member 1) Why are we spending our time on the performance reports that are just showing historical trends, when we know from the ad hoc work that we can be more useful predicting and helping the police force to prevent crime? (Team member 2) There is some really interesting research on data analytics in policing, and the potential for things like predictive algorithms, big data analysis and artificial intelligence to make a big difference. Sadly all our police force seems interested in old fashioned trend line reports and historical analysis. I have actually been looking at jobs in other organisations that are more forward looking – I know it would leave you short staffed, but I want to work for a more forward looking organisation. (Team leader)
She also spends some time talking to people who request and use the ad-hoc reports produced by her team, and others in similar roles who do not. She notes the following feedback. The report you did last year on young people who repeatedly go missing from children’s homes was really useful. I was able to highlight the homes that saw this as a police issue and just kept reporting children missing, and see if we could work with them to help prevent these troubled and vulnerable children going missing in the first place. In the three highest reporting homes, we have now reduced the number of calls to us by around 40% (Neighbourhood Policing Inspector A). I usually know what the issue is. I have 24 years of experience in this area of policing, and sometimes your reports don’t help me as they try and say the real issue is something else. My professional judgement tells me that there is not a major drugs issue in my area, and the rise in thefts, burglaries and shoplifting is not related to drugs, but your team seem to think differently. (Neighbourhood Policing Inspector B). You did us a report on repeat callers to the police emergency line. We were able to identify the top 40 repeat callers, and are starting to work with other agencies such as the mental health teams and social services to identify the root causes of these people calling the police. I am hopeful that we can reduce the demand from these people on the police as well as get them the help they need. One person called us over 100 times in a three month period, but since last month they are getting some mental health support, and have only called a couple of times since. (Head of Emergency Call Handling) We need to be much better at identifying and using evidence of what works to drive improvements in our service. We are not very good at collecting evidence of what works and what doesn’t and evaluating this to help us make better decisions. I know many other police forces have teams who use the data they gather from initiatives to evaluate police tactics, and this has allowed them to share best practice across the force. We should be able to do better. (Senior manager and member of the Resourcing Board) I don’t really look at the performance reports, unless the line is drastically pointing downwards. Because if it is then my boss will get on to me about fixing it. Policing is more about professional judgement, and while we have to produce these reports, they are not really that useful to be honest. (Neighbourhood Policing Inspector C).
Conclusion
Organisations, including law enforcement, face an unprecedented explosion of big data due in part to the widespread use of social media platforms in society and digitalisation initiatives. It is argued that big data are ‘worthless’ in a vacuum, and, thus, only when it is leveraged to drive organisational decision-making, does its potential usefulness and value become apparent. Consequently, organisational require effective methods to transform large volumes of rapidly changing and diverse data into insightful conclusions in order to support such evidence-based decision-making. Despite this, data analytics have created several challenges in the world of policing.
Footnotes
Declaration of conflicting interests
Funding
ORCID iDs
