Data Science in Petroleum Engineering

1. (a) ‘Playing around with data’ Vs (B) ‘Trying to investigate the complex interplay between reservoir geology, drainage principles of oil & gas, reservoir fluid dynamics, reservoir thermodynamics and then translating into an equivalent math’: What will happen if (a) is directly handled by budding petroleum engineers – in the absence of undergoing enough training and exposure in (b)?

In the absence of (b), whether the data-driven person will be able to understand the complex fluid flow driven towards the production wells?

2. While it is feasible to deduce various kinds of insights while analyzing the data, will it really be useful to make sensible and economic business decisions – in the absence of (b)?

3. By what means, the details gathered by a cloud based on the way that you used to deduce various kinds of data (core/log/well/PVT data) would remain useful in a petroleum industry?

The concept of database itself projects that we have non-uniform set of data associated with various scales. If so, how could we transform the data efficiently (addressing the scale issues) – despite cleaning the data precisely?

How meaningful would be – the concept of ETL (Extracting from cloud; Transforming the data & Loading into a database) remain appropriate in characterizing a petroleum reservoir?

Now, how will we be able to justify the data to remain to be ‘usable’ by a reservoir or petroleum engineer – in the absence of validating all the related cleaned data?

Are the data engineers (involved in collecting and building the data and thereby ingest data to the cloud and then, finally to the data base) have a sound knowledge of rock and fluid properties associated with various scales?

4. If usable data from the database is directly analyzed, then, will we be able to recommend critical EOR technique decisions – based on the insights?

On the other hand, if budding petroleum engineers use these data to build deep or machine learning algorithms, then, how will the budding petroleum engineers be able to deduce and appreciate proper insights?

In the absence of experienced industrial experts, how could a data analyst such as a budding petroleum engineer – would be able to make data-driven decisions (which is going to be happen in the future) – as a function of the past ‘data’ that has already taken place?

Whether the concept of ‘Data Visualization’ could be justified by a budding petroleum engineer? How would he/she be able to visualize the data in a proper way – in the absence of an in depth ‘domain knowledge’?

Would it remain feasible to project the oil/gas productions – from the deduced patterns – associated with the data – given the time and space dependent rock and fluid properties?

Despite gaining a significant statistical knowledge, by what means, will we be able to correlate a recommendation system and a ranking system with that of a reservoir heterogeneity?

Feasible to train a model in the absence of (b) – despite having an expertise in coupling statistics and computer science?