Speaking at the 2013 IRM-UK Enterprise data governance and business intelligence conference in London last month, in what was billed as a joint presentation with PDO, Walid el Abed (Global Data Excellence) described data governance as a means of enabling ‘value driven leadership.’ In drilling and production, it can be hard to agree on key metrics. But it is necessary to settle on some tangible business objectives. El Abed advocates using a ‘key value indicator’ to judge overall data quality and usefulness. This can be rolled-up from data quality and volume metrics. Typical KVIs for the upstream are lost and deferred production and non productive time. When using past data to predict the future, el Abed recommends working from a ‘near real time through tomorrow’ time-frame to build a data predictivity time table.
Ian Pestell (Cisco unit Composite Software) reported that using extract transform load (ETL) technology to load the data warehouse is ‘not the best agile approach.’ For one client, adding a single column took six months! Moreover, half of all data warehouse projects fail and, where they do work, around 70% of data is ‘dark’ i.e. is never queried. The conventional approach suffers when there are changes to operating systems. It is also fails to incorporate many external data sources like weather and social media feeds. By combining data virtualization (thanks to Composite) along with Cisco’s query optimization technology a truly agile data environment is possible.
Lee Edwards of Data Management International (Dama) described how to ‘thrive’ in global data management. This involves combating many silos—cultural, budgetary and physical. The cultural silo is defended with the ‘not invented here’ approach and the fact that all use the same words to mean different things. Dama’s approach involves listening to and observing users, looking out for common themes. Other tricks include anticipating requirements and delivering a first cut data model early. Later key data sources can be identified, refined and embedded in a preliminary physical model. Edwards advises repeating the questions ‘how do you know?’ and ‘what is it you need to understand?’ This is because ‘business are good at hiding their intelligence.’
Exhibitor Information Builders told us that they have been developing business intelligence solutions for Petrobras, Vale and Upstream Professionals. Tableau Software’s visual analytics although not industry specific, is used in oil and gas. Mark Logic’s application underpins BP Trading’s ‘Where’s my ship?’ application for tracking vessels along with weather data and tweets. More from IRM-UK.
© Oil IT Journal - all rights reserved.