Call

1st Open Call

Challenge

EDI-2018-2-UBIMET. Quality control of weather observation data

Project Description

Modio will develop a novel methodology based on i) tensor theory II) big data analytics frameworks (e.g. Ignite) to detect (flag) false observational data indicating stations that produce measurements of questionable quality. The crux of our methodology is its capability of detecting anomalies from the time-series of measurements that correspond to a group of consistent stations with respect to the following dimensions: (1) time and measurements, (2) space and measurements, and (3) time and space and measurements. For example, it will be able to detect an anomaly when a station reports a series of unexpectedly high temperature values, despite the fact that it is physically close to other stations and produces similar values with these stations in the other attributes at similar time points.

Discover the product’s business and technical approach:

Company description

Modio Computing PC (www.modio.io) specialises in machine learning, data analytics, applied statistics and large scale processing and messaging for the telecommunications and the Internet of Things sectors. Modio has several years of experience in telecommunication networks, large-scale distributed systems, messaging, statistics, machine learning and analytics and has delivered projects in telecommunication companies, security companies and advertisement companies.  Commercial projects involve our expertise in messaging, cloud provisioning, mobile media, network management, machine learning, data extraction and event aggregation and rule-based event processing. Modio’s flagship product is the Qiqbus platform for streaming analytics, www.modio.io/qiqbus.