It is true that the most of the research work on predicting natural disasters has not been dependent on big data tools. Researchers, for decades, have employed whatever statistical tools available and have relentlessly followed their pursuit. But what cannot be denied is the fact that big data has created a giant leap in the analytical capabilities on large data sets. And more than that, with access to such tools now available in every hand, startups too are able to innovate on research subjects that were once considered accessible only in academic computational labs. With a devastating earthquake killing more than 5,000 people in Nepal last week, I thought it is an opportune time to review how far are we from reaching a stage when loss of lives from natural disasters would be a thing of the past. And expectedly, startups are leading the way here too. Although it might not be prudent to speak about financial losses in the same breath as loss of lives, one must admit that participants from industry, government, insurers and even trading firms are as keen for such solutions as the common man. And luckily, their commercial interests are playing a good role in driving funding and innovation in this stream. Today, I’ve talked about three such data analytics startups, each addressing a different form of natural disaster prediction (earthquakes, volcanic eruption and pandemics).
The first startup that impresses with its innovation is the Jersey based Terra Seismic. If you visit the home page of this 2012 founded venture, you will see a crawler that boasts about the earthquakes and tsunamis it predicted accurately (1-30 days in advance) over the past few months. Let me add that predicting earthquakes is certainly not an breakthrough innovation. Successful earthquake predictions began in 1975 in China and, in the recent times, have been relatively more successfully employed by the Japanese. Methods might differ. The original Chinese effort focused purely on sustained seismic activity to build predictive models while the Japanese rely upon an extensive network on sensors deployed beneath the surface to alarm them when an earthquake triggers beneath. The latter is extremely expensive and cannot alarm the population beyond a few minutes in advance, and the former is non-reliable as seismic fault zones are almost active across tectonic plates. And this is where Terra Seismic and its accuracy-enhancing big data analytics smells like ‘hope’. Employing python-coded open source software and running extensive satellite and sensor data on Apache web servers, Terra Seismic’s algorithms keep a tab on seismic activity in sensitive regions and keep building factors-based earthquake prediction models from historical data. The idea is simple. With each iteration of data set, the accuracy will deepen. It is not that such earthquake big data research is only being carried out by Terra Seismic, but already armed with the prowess of 90% accuracy in prediction and an aggressive pursuit of roping in commercial clients for its services, the effort is on its side.
V-ADAPT is a software based risk assessment startup for the transport and logistics industry. What stands it out from others is that it specializes in airborne hazards, particularly volcanic eruptions. And more than that, its claim that its services are ‘near real-time’ adds a huge impetus to its utility. To its customers, V-ADAPT provides near real-time access to satellite data, along with the software tools that keep the users informed about current volcanic activity in the form of a layered interface for data interpretation. Its service portfolio also includes ash cloud detection and updates on thermal conditions. The premise is to use a remote sensing data detection system that records changes in a volcano’s surface temperature preceding eruption activity. V-ADAPT rounds up its service offerings with a volcanic ash cloud tracking model and forecasts on ash cloud movement based on simulations made before the eruptive event. This can be extremely vital in predicting how the ash cloud would adversely impact flight paths, nearby habitations or industrial activity locations. V-ADAPT was incubated with the help of Nanook Tech Ventures, a startup accelerator founded in April 2013, employs UAF-developed technology. Last year, V-ADAPT earned a high profile client in the South Korean government.
Founded in 2008, Metabiota is a San Francisco based startup specializing in pandemic detection, discovery and outbreak investigation. It calls itself a pandemic threat management, mitigation and preparedness service provider that uses health data analytics. It gathers its data from multiple sources, including viral discovery in the field, anthropological research in disease hotspots to identify how viruses cross from wildlife to humans, and tracking social media trends. Metabiota is led by doctors, medical researchers, and anthropologists and primarily uses big data to seek out emerging diseases among animals in hopes of preventing their spread among humans. Its clientele includes USAID too. The company raised $2.38 million of funding in 2014. It also has a relatively well-known non-profit arm with the name Global Viral Forecasting Inc. (GVF).