Investigative research: Definition & Process
Digging up relevant information about a subject from a collection of data is a skill required in many fields. From investigative journalism to criminal investigation, finding a coherent story from seemingly unrelated data plays a considerable role in many activities. And this is where investigative research comes into play. Let’s have a look at what it is.
What is investigative research?
Investigative research is used to describe various methods and tools for finding information about a specific subject or incident which may not be apparent. The process also includes searching for relevant data which may not be readily available. Investigative research is used for everything from thwarting terrorist activities to finding money laundering. In a business setting, these tools may help you do market research, understand bottlenecks, or anything affecting the revenue stream.
In the age of information, the nature of research has changed quite a lot. Be it academic papers on cancer or an article about data analysis, there’s no shortage of available data. The difference is going through all of this data to find the information we need.
Most of the information is unstructured (up to 90%), in the form of emails, chats, or other formats, and traditional tools can’t handle this. Another major roadblock is the data silos that exist in many organizations. It’s tough to connect data that exists in many sources and build a clear picture. In a survey conducted across eight countries with 131 decision-makers, 85% of respondents found it difficult to access information.
Let’s discuss tools that can help decision-makers conduct investigative research efficiently and effectively.
Major tools for data investigation
Researchers have plenty of tools and resources available for diving deep into the data and unearthing hidden facts and stories. Around 21% of decision-makers in IT, from the above study, wanted IT solutions that would let their team work on higher-value assignments. They wanted to automate more straightforward or repetitive tasks.
Here are a few tools that will help an investigative team do just that.
If you’re investigating with your data, this may not be entirely relevant for you. But even if the organization has plenty of data, you may need to connect with other data sources, such as the value of currencies on a specific date, to come to conclusions.
Doing a simple Google search may give you a lot of data, but you may have to dive a bit deeper if you want more. These tools can help you here.
Google Public Data Explorer
This tool is entirely free to use and works more or less like google search. You can search across a wide range of data sets available publicly. Google Public Data Explorer also has good visualizations and is highly intuitive to use. You can also upload your datasets for easy visualization.
Dataportals.org is a list of websites where you can access public data. Most of them are government websites or data portals, but you can also find websites with data related to businesses and other fields. The website is free to use.
Data analysis (or finding the information from the data)
Once you have the data, the next thing is to find patterns and trends from this. While structured data is relatively easy to analyze and find patterns in, things get complicated with unstructured data. And a huge chunk of data, particularly in criminal investigations or audits, is unstructured.
This means going through tons of emails, chats, other communications, and more to find what you’re looking for. The good thing is that we now have the tools for that.
Sinequa Enterprise Search
Sinequa’s enterprise search solution can help you significantly reduce the resources for investigative research. It comes with hundreds of pre-built connectors to data sources. Sinequa can help you quickly find what you’re looking for from tons of unstructured data. The system uses cutting-edge AI and natural language processing to derive answers to your questions. Recently a leading financial intelligence unit successfully used Sinequa’s solution to counter money laundering activities.
Sinequa can help you quickly find what you’re looking for from tons of unstructured data. The system uses cutting-edge AI and natural language processing to derive answers to your questions. Recently a leading financial intelligence unit successfully used Sinequa’s solution to counter money laundering activities.
Sinequa’s enterprise search can significantly accelerate investigative research. During most investigations, data comes from multiple sources. By presenting a complete view of every case or subject under investigation, Sinequa helps researchers quickly reach their conclusions.
By combining Natural Language Processing and Machine Learning, Sinequa offers a constantly adapting strategy for the changing scenario. For example, in the case of financial crimes, the perpetrators are constantly coming up with methods to overcome rule-based solutions. But the Sinequa system is continually learning from new data sets. This makes it capable of finding fraudulent transactions that would otherwise go undetected or take significant resources to prevent.
Investigation processes vary according to the specific type of investigation. The process for a criminal investigation won’t be the same as an investigation into unexpected revenue loss. But we can develop a general process flow for finding answers to a question.
1. Define the question
This is the first step in any investigation. Before the investigation proceeds further, you have to define the entire scope of the investigation. What do you need to find, where are you going to look, what are the resources available, who are the stakeholders, etc. Of course, sometimes, new information that broadens the scope may reveal itself throughout an investigation.
2. Collect the data
This is where you collect all the data about the matter. This could be telephone records, chats, emails, credit card histories, bank statements, and more. The nature of the investigation will dictate the type of collected data. In some cases, the evidence may be purely digital, while in some, it could include DNA, witness testimonies, and other forms of data.
3. Filter the noise
In a corpus of data, there will be a lot of noise. For example, let’s say you have all the emails between employees of an organization to investigate defrauding the shareholders. The collection will have birthday and anniversary greetings, gossip between the employees, spam emails, and more, which have nothing to do with your investigation. Same goes for call records, bank statements, or any data collection.
You can also filter out the data based on the time, and in some cases, even based on geographic locations. Particularly in criminal investigations, location-based filters can uncover critical insights.
After all the filters are applied, you may validate a subset of raw data to make it a bit more authoritative and reliable.
4. Find the connections
Find how different entities are connected, any communication between them, and determine which entities are involved in the case. For example, if you’re investigating the call records or communications of a terrorist cell, try to figure how they’re connected, how and what they’re planning, and who are all connected to them.
If you’re trying to figure out a conspiracy to commit financial crimes, connect who all are communicating with whom and what they’re planning to do. The frequency and content of communication can tell you these relationships.
5. Expand on the connections and build a timeline of events
Build up a timeline of how everything started, showcasing every turning point. A timeline will give everyone a clear picture of what happened. It will also help you figure out what’s missing in your investigation.