The SigmaIQ methodology: An information quality perspective on oil data

Information quality (IQ) theories and frameworks have been increasingly studied and applied in various organizations to assess, improve and monitor the quality of their information products. Yet information quality problems remain pervasive in many organizations, industries and government institutions. For example various governmental as well as non-governmental organizations, institutions, and companies collect, compile and distribute information products in order to satisfy the needs of information consumers in the energy industry. What are the qualities of their information products? How do they collect or disseminate their data products? Why do information quality problems exist and how are the problems created? How can we assess the quality of an information product in the energy sector? What can be done to address the quality issues so that data consumers can trust that the data they are relying on in decision making are of the highest quality? As an important case of the pervasiveness of information quality issues in the energy industry, this dissertation investigates IQ problems in crude oil exploration and production. Specifically, this dissertation focuses on two main problems.;1). The difficulties of estimating total world oil reserves by assessing the information quality of oil reserve data from various sources .;While high quality data can help oil companies and governments reduce risks to investment and company bottom lines, bad data can lead to economic loss. In this research, I investigate a framework for assessing the information quality (IQ) of world crude oil reserves. The framework is applied by calculating the information quality ratings of several well-known information sources. The information quality dimensions assessed in this framework include Data Decay, Data Integrity and Reputation, Data Coverage and Completeness, Degree of Compliance with Data Standards, Expertise of Data Source, and Degree to which Data was Vetted. Using the proposed framework, I quantified the information qualities of the sources on these IQ dimensions and calculated combined information qualities, or SigmaIQs, for each data source, ranking the sources accordingly. The SigmaIQ assessment results were used to calculate a composite estimate for world oil reserves. The distribution of this estimate is discussed in the cases of uniform, triangular, and Gaussian distributions keyed to the information qualities of the individual sources. The robustness of the proposed framework was tested by weighting the IQ dimensions in multiple pseudorandom ways and calculating weighted SigmaIQs for each pseudorandom weight scenario. The resulting rankings of data sources in terms of their SigmaIQs were compared to the original ranking.;The results show that applying different weights for individual IQ dimensions can result in different SigmaIQs for the information product. This can be significant. In some cases the ranking of the sources based on their SigmaIQs have also changed when different weight vectors were applied. However, the overall trend of ranking did not change greatly, for example, the higher ranked sources remained relatively higher in rank, and lower ranked sources remained relatively lower in rank despite the different IQ weights being applied. The IQ-weighed world oil reserve based on seven data sources was calculated to be 1386 billion barrels (bbls).;The results and conclusions of this research provide insight and guidance to users not only in decision making using data products from the sources, but also in applying the framework to similar problems in the energy domain.;2). The problem of predicting the Hubbert Peak of world oil production by assessing the information quality of existing prediction data from various sources.;A similar IQ framework as that developed to address problem 1 was applied to assess the SigmaIQs of world peak oil production forecasts from several well-known sources. The assessment results were then used to calculate IQ-weighted dates for the oil production Hubbert Peak (that is, the date of peak oil). The IQ-weighed Hubbert Peak date calculated using predictions from 7 data sources illustrate a range from as early as 2016 to as late as 2026 for the date of world peak oil.