In * Gather business requirements before gathering data and define a clear objective
* Evaluate data requirements. This process usually requires inputs from all business stakeholders; analyse and ascertain what data is crucial and what is not, for any given use-case
* Big data is never an IT driven project; identify a business sponsor for the project
* Big data is required to solve business problems
* Allow data scientists room to breathe to construct their data experiments and prototypes using their preferred languages and programming environments. Then, after successfully executing such POCs, these can be operationalised in production using altogether more professional IT methods, modifying and optimising as required
* Avoid data duplications as much as possible; when it comes to big data, every redundant duplication incurs an incremental cost
* Sandboxes for experimenting must not be frowned upon because while digging for useful/actionable insights in big data, we often dont know a-priori what it is that we are looking for until we find it; allow for experiments without a clear sense of direction
* IT should be ready with data sets and recommended solutions once business groups come up with their needs
* IT should make big data a self-sustainable model where additional investment is not required to generate big data
* IT managers need to build special capabilities collaborating with enterprise mobility to enable quicker data compiling methods
* IT teams need to take a holistic approach to big data pondering over how to evolve their core IT infrastructure to facilitate large big data projects.For instance, discussions with IT heads on how to approach delivery of a big data reservoir would be beneficial
Add new comment