With the help of Machine Learning and AI, the company makes sense of data, identifies what’s relevant and what’s not, and provides actionable insights that enable clients to make informed decisions
Geospark Analytics, a Washington DC-based startup, was founded in 2017 with the purpose of helping organizations avoid risks and make decisions based on highly accurate data. The company believes that real-time data can be harnessed and categorized to make a significant difference in analytics, risk assessment, and decision-making. Geospark Analytics uses the power of spatial data to identify what’s relevant and provides insights beyond simple visualizations of locations on a map.
In the age of data profusion, it gets difficult to accurately understand the real issue and make an effective decision. The company gathers a constant flow of exponential volumes of data through a combination of neural networks, AI, and predictive models to provide meaningful insights to its customers.
Merging location intelligence with Deep Learning and AI can lead to more precise results in the future. The company geotags all the data that it gathers, categorizes it using Machine Learning models, and using the data, assesses activity and stability levels.
Geospark Analytics primarily caters to the threat and risk assessment market. Most of its clients are in three verticals: Government, NGOs and commercial. The government clients include foreign affairs, defense and intelligence organizations. The NGO clients are quite varied, ranging from those looking at investments for development to those working for disease eradication. While the company’s commercial clients include those who provide security to Fortune 100 companies and travel safety companies. It is now exploring opportunities in the financial market such as hedge funds and reinsurance.
The company has a dedicated global threat and risk assessment platform called Hyperion that is designed for better decision-making. The platform provides a simple, intuitive event discovery and monitoring capability by identifying and forecasting emerging events on a global scale. Its AI analyzes streaming content, assigns stability scores, and makes forecasts of stability for every spot on the globe. These assessments can mitigate risk, recognize threats, and identify opportunities.
Geospark Analytics recently signed an agreement with Twitter enhancing its Hyperion platform with hyper-focused real-time breaking events, further strengthening its Machine Learning and AI models.
“A simple stream of real-time data is overwhelming and ultimately drives people to information overload and decision-making paralysis. The age we live in is characterized by Infobesity, an emerging condition that looks at information overload as the root cause of the inability in effective decision-making”, says John Goolgasian, Chief Operating Officer, GeoSpark Analytics, in an interview with Geospatial World.
What sources of information do you rely on, and how do you ensure that the information is authentic?
Hyperion provides advanced AI capabilities within an easy-to-use platform that delivers real-time insights and automated monitoring, alerting, and forecasting of events that could pose a risk to people, places, and investments. We are proud of the disruptive technology that we have created and our platform is only going to get smarter with time. The way people monitored risks and threats in the past are outdated and unsustainable.
Hyperion integrates both long-term and breaking information from structured and unstructured data from news media, social media, natural disaster information, travel warnings, economic indications, and others. In the first six months of 2019, Hyperion has processed data from over 6.8 million sources averaging between 70,000 and 100,000 individual pieces of relevant per day in over 100 languages, all auto-translatable. We actively work to avoid bias in the system. Though we keep the provenance of all data, our users make the final judgment call on the value of individual pieces of information to their mission. However, we are confident in the aggregate assessments that our platform makes, as our carefully refined models have proven to be leading indicators for falling and rising stability trends.
Tell us about the functioning of Hyperion, and how do you power it with Machine Learning and AI?
Hyperion uses AI and Machine Learning to compile automated risk and threat intelligence information by ingesting massive amounts of data, determining what is relevant, and delivering the user streaming assessments and alerts. Leveraging Machine Learning ensures that Hyperion is the most efficient and scalable solution for risk and threat assessment.
What sets Hyperion apart is the innovative, cutting-edge technology that drives it. While there are a variety of threat assessment solutions available in the market today, what differentiates Hyperion is how its global threat and risk assessment is done.
How do you monitor millions of live updates and feeds simultaneously?
This is exactly the problem we worked to solve. It is impossible for any human to monitor millions of live updates and information feeds simultaneously. That is why we have worked hard to create a powerful model that automates the monitoring, assessment, and forecasting of breaking events and global stability patterns for our users – we have solved the human scale problem. By leveraging Machine Learning and AI, we efficiently and effectively make sense of an infinite amount of information. We allow the machines to do the work, so the humans can focus on what actions need to be taken.
Could you tell us a bit about your Stability Index?
We measure stability with our Pulse models. Pulse calculates long-term and short-term stability changes using factors from data that is constantly streaming and normalized by long-term indicators, producing an extremely accurate view of the world. The model continuously assesses levels of activity and risk, defines normalcy, and combines them into a composite score of stability for every spot on the globe. Pulse allows you to rapidly review stability in every country, 7,000 regions, and 1,200 cities covering every spot on the globe in real-time. This is a unique AI technology that is amazingly accurate.
What is the most efficient way to identify baselines of global activity patterns and define micro-activity trends?
The best way is by using Machine Learning. With our Machine Learning models, we can track relevant information as it breaks. But more importantly, we have baselined the world’s activity and built a living stability model of the world that updates thousands of times a day to identify micro-stability changes; enabling our clients to act before it’s too late. Hyperion is identifying low-burning issues that turn into major events, and it identifies where these issues are having small effects on stability prior to them turning into major problems.