Use of data science for most organizations is not a discretionary topic – it is vital to their success and competitive advantage in many areas. This is particularly true in the context of investigations, cyber risk and compliance solutions and other legal matter. Only firms with strong legal-domain knowledge coupled with expert-level capabilities in electronic evidence and data science are able to successfully manage a matter or solution in a timely and cost-effective manner. Use of statistics, unsupervised learning, predictive analytics and other data mining tools are leveraged to find the gold otherwise lost in data.
Working with our client and team, we leverage a 3-step, sequential process for all matters and consultations involving electronic data:
- First, determine all relevant legal, policy, compliance and regulatory issues as well as client priorities and goals.
- Establish and identify the electronic data necessary to accommodate goals in step #1.
- Then, and only after completion of step #2, determine if and how data-science tools may add value – very often, this will be the case.
This 3-step process is the only way that cases involving electronic data and data science can be handled. In our experience, any other approach leads to misapplication of data-science tools.
Advanced Data Analytics
There are often occasions where boilerplate, pre-configured data-analytics tools do not provide the fine-grained level of control needed to truly glean as much useful information as possible from data. Accordingly, we use API’s offered by state-of-the-art open-source and commercial applications to tailor a solution specifically to the needs at hand. We find that the following technologies are most useful in investigations and compliance solutions to find and utilize the most critical information quickly and efficiently. These tools are often used together and in an iterative fashion to maximize results:
Information Retrieval
Makes use of search technologies to identify relevant information in millions of records across very diverse data sets, e.g. emails, spreadsheets, social media, business records, web data, database records, etc.
Information Extraction
Uses machine-learning and natural-language processing techniques to extract previously unidentified people, locations, associations, entities, account numbers and nearly one thousand other entity-types from data. Similar techniques extract keywords, concepts, sentiments and other “emotions” to help narrow data to particular topics, events or groups.
Graph Databases and Visualization
Finds key relationships, sequences and anomalies across transactional data, social media, text files and other structured or unstructured data.
Timeline / Time Series Analysis
Fleshes out and refines chronologies of events to help tell the story behind the truth.
Geolocation Analysis
Brings context to the locations of events and entities as gleaned from large numbers of records.
Correlation Analysis
Establishes correlations among millions of facts to drill down on possible causes and relationships.
Preventive, Proactive Services
“An ounce of prevention is worth a pound of cure!”
Patterns, facts, statistical information and other metrics discovered during our initial investigations can be used directly in predictive models within an enterprise infrastructure to detect, prevent and mitigate unwanted activity. We are uniquely qualified to provide this type of consulting and advice.
Licensed private investigator in CA (#18413) and IL (#115.02338)