Predictive Analytics for Risk Avoidance

点击翻译按钮获取中文摘要

The aim of the project is to establish an architecture, or framework structure, to allow for searching and mining datasets of safety related events in the chemical and oil and gas industrial infrastructures. The tools of predictive analytic will then be applied to extract information on potential averse identified situations and/or probable states in future operations of the infrastructures. Based on the extracted information it might be possible to implement, modify, operations, or take corrective actions in such a manner that the predicted adverse events are potentially prevented or avoided. The project will focus on identifying, defining and transforming data into a common structured data formats coming from all different data sources in the chemical and oil and gas extraction industries. Data analysis techniques will focus on building algorithms, and other probabilistic methodologies to perform predictive analytic on the data that is obtained and integrated for benchmarking from various sources. Useful work should then be possible, allowing to query data, create summaries, distributions, trends, projections, predictive probabilities, reports and charts of the integrated available data. To quote Ketan Puri of Infosys: "The way the oil and gas industry works with data can be depicted as a funnel, with raw data going in one end and usable data coming out of the other". In this project the correct interpretation of data will increase certainty in decision making of safety managers / professionals resulting in safer operations of the infrastructures. The relevance of the work for the company is that the software developed will enable the company to offer a service for industrial facilities that will predict the potential occurrence of adverse events at the facilities with a probability attached.