Measuring your ideas: In this step we formalize your ideas, the nature of the end product, and what you would like to learn from the endeavor. The hardest part of any project is framing it properly.  Once framed, much of what follows flows naturally.  Important to this task is the observation that all successful analytic endeavors are collaborative. 

You know your business better than anyone else.  You know your core principles and aims.  We provide the expertise to turn these into measurable quantities. To begin this process of measurement, we start with two simple, related questions.  If you could have what you wanted, what would it be?  If you could have what you wanted, how would you change your behavior?  Through successive meetings we learn your business rules, the form of the end product, build project statements, illustrative examples, and prototypes.  This refines your project goals first into clearly articulated statements and then explicit measurements of all quantities of interest.

Refining the raw asset: Second only to problem formulation is ensuring the accuracy of your data.  As an integral component of our modeling process, we conduct a thorough and meticulous review of your data.  There is always a delta between what one expects to see in one’s database versus what is actually there. 

An initial review of your data will produce a report detailing any errors, discrepancies, and inconsistencies.  To aid your data team in resolving this, this report will contain examples of the above, code to locate them, and recommendations on how to attend to them for modeling purposes.  These recommendations will form the basis for collaboration with your subject matter experts on how to ensure that the data utilized in our analysis meets the highest standards of quality and integrity.

Finding the relationships: With the project aims properly formulated and the data in place, we move to find the relationships in your data.  Our modeling process is the result of ten years of research combining results from the giants of advanced mathematics, the latest techniques, and the most powerful computers.  All of this is put into the service of creating models which are precise, interpretable, and incorporate your business rules.

Sharing the insights: All models are constructed to inform the action of the stakeholder.  The deliverable from Model Construction is the first formalization of the relationship(s) among the quantities of interest.  These relationships will be detailed in a report which includes all assumptions made during the modeling. 

We will provide comprehensive guidance on the significance of these models and their implications for your business operations.  Presenting this to the client often inspires additional questions, refined hypotheses, and additional sources of data.  Upon receipt, we will perform the additional Data Due Diligence and further refine the model.  This proceeds iteratively until all questions have been answered and sources of data exhausted.  In particular, the client will know what is driving the relationships of interest and what is superfluous.

Turning insights into action: Finally, we transform the results of the analysis into the end product, whether it be a spreadsheet, a stored query, or a function contained in your transaction processing system.  We will collaborate closely with your infrastructure team to ensure that the deployed model is functioning optimally, meeting all performance standards, and seamlessly integrating with your existing systems. This often consists of a few rounds of testing each phase of the technical workflow.  This collaborative effort will guarantee the model’s reliability and effectiveness within your operational framework.