A changing landscape
Expectations around analytics platforms have changed. In the past, analytics was about reviewing data, comparing it to previous events so that historical reports could be produced. There is now a desire to ask more sophisticated and interactive questions of the data. Organisations want more. They have dashboard fatigue from the old approaches. Analytic tools are now expected to invite the next question of the data, to spot non-obvious relationships, and to get into an interactive dialogue with the user to solve a specific problem.
The range of users working with analytics platforms has changed as well. Previously, analytics had to be simplified so that high-level business people could use it. Today, analytics has several audiences. This could include leadership, general business users, domain-specific analysts, and highly technical data scientists. The expectation is that your platform should be able to cater for several job functions.
The demand for visualisation
Because of this change in expectation, and because more complex questions are being asked, sophisticated visualisation has become hugely important. Knowledge representation has been researched for many years. If it is usable enough, it enables an ongoing dialogue with the user about the data. No longer are we representing data through pie charts or histograms alone: we’re now looking at how entities of real meaning are related to each other and what the consequences of those relationships are.
The historical challenge has always been to visually represent data in a usable fashion. Today, we can highlight visually relevant facts and relationships, which is powerful. Previously, screens would be crowded with visuals which provided no insight. Usability is catching up with potential, and the knowledge graph can help tell the story the data has to tell. You no longer need to think like a programmer to gain insight. You just have to just think about the problem.
As we look to develop new innovations in analytics platforms, I believe AI will play a pivotal role. We’ve reached a point where we have a lot of information and data, and the analyst needs help. The help comes from AI, and I like the term used by Gartner, which is “augmented analytics”.
A recent Gartner report on Analytics and Data Science stated: “Augmented analytics is the future of analytics. The proliferation of augmented analytics causes the collision of analytics and business intelligence and DSML products. The converged capabilities will enable a more empowered class of analytics consumer — the augmented consumer — to add more business value by exploring their own data.”
Today, organisations have huge amounts of data, and we need the machines to help. We don’t need the machines to take over the decision-making process, but we need them to augment the analyst. What I mean by this is:
- spotting anomalies in the data;
- providing machine-driven alerts;
- helping knowledge representation through approaches like topic clustering;
- pointing out non-obvious relations through entity resolution;
- giving us screen prompts.
It’s about not depending on the analyst to know what question to ask and at what time, the machine must prompt the analyst to go in a certain direction. However, the analyst remains in control of what decisions are made and the consequences of those decisions.
Having said all that, we’re still at an early stage. There is a disconnect. Within an organisation, analytics is often separate from machine learning and data science groups. However, we are now seeing more convergence. For a long time, AI has been a solution looking for a problem. Whereas now, analytics has developed to a sufficient extent to be able to take advantage of AI. At Siren, it’s about the user experience for the data scientist, analyst, divisional and overall leadership. If they can get value through the AI, in all its flavours, it’s no longer an academic exercise.
Customer influence on platform development
Our customers are looking to push the boundaries on data volume and scale in terms of investigative intelligence. It’s not unusual to have many billions of records as part of an investigation. Which begs the question what’s the art of the possible and how to separate vendors that can operate at that scale from those that can’t.
It’s also worth mentioning justification around decisions. For example, when analytic findings are used in court:
- law enforcement needs to be clear about how decisions were made;
- how the investigation was run;
- what was the logical justification for matching two records;
- was AI used to form an opinion and was it biased?
There is a requirement for an audited case history of all these processes. In terms of developing our platform, many clients are looking for integrated case management so that they can present the whole picture of an investigation’s workflow.
Data governance is another important consideration. If you are dealing with classified information, who has the right level of access? Who is able to grant access? This goes beyond just security controls, into the overall stewardship of the data. Who is looking at data, and why?
Law enforcement and digitalisation
Society is now digital. COVID-19 has forced organisations to digitally transform. Therefore, if most of our lives are lived online then crime tends to follow such as online fraud, cyber-attacks, identify theft and human trafficking. All of these crimes leave a large digital footprint. Unfortunately, in terms of law enforcement, they are often playing catch up.
After years of under-investment in technology, law enforcement are now looking to bypass a couple of generations of technology. Let’s not go through the pain of what was done in the past, let’s just go straight to what is needed today to run efficient data-based investigations. For platforms, AI needs to be integrated into the analytics problem to ensure the right answers are found.
Siren provides the leading Investigative Intelligence platform to some of the world’s largest and most complex organisations for Investigative Intelligence on their data. Rooted in academic R&D in information retrieval, distributed computing and knowledge representation, the Siren platform provides integrated investigative intelligence combining previously disconnected capability of search, business intelligence, link analysis and big data operational logging and alerting.
Among Siren awards are Technology Innovation of the Year and the Irish Startup of the Year (Ireland’s National Tech Excellence awards). In 2020, Siren was named as a Gartner Cool Vendor in an Analytics and Data Science Report. For more information, visit www.siren.io.