Science4Data Solutions

Science4Data R.E.A.L. Approach

We have a four staged value creation approach that is derived from our scientific practice applied to exploiting business opportunities created through information processing. The hypothesis which states some view of the real world conditions. The data signals that provide evidence for the hypothesis of how the real world is functioning. The computational processing using the mathematics derive artifacts that can be reviewed by experts to validate the hypothesis. The learning loop where computational and human expertise are fused to visualize the insights that improve the models so that they are closer to reality. This approach has a re-enforcement mechanism that creates a continuous improvement cycle which establishes a superior competitive advantage for our client readiness to dominate their industry sector.

Real World modeling through knowledge representation

Evidence Signal Data Capture and transformation

Applying Insight Science engine processing

Loop Learning from visualizing insight signals and adjusting the models

We engage with our clients using the CMU Software Engineering Institute IDEAL framework for process improvement. This framework is similar to Six Sigma which is well familiar to corporations that have implemented best practice innovation change management. The IDEAL framework is focused on opportunity discovery which is complementary to the typical Six Sigma practice focus on cost reduction through quality control. We have reinterpreted the IDEAL framework for insight science.

The IDEAL acronym staged activities are:

Initiating

• Understand the context of the potential business value opportunity.
• Identify all stakeholders and indirect (customer) beneficiaries
• The initial infrastructure scope, cost and risks assuming success in realizing the opportunity.

Diagnosing

• Profiling description of the current and desired future states of the product/service or artifacts.
• Consider the alternative opportunities and solution paths.
• Formulate the recommendations.

Establishing

• Data experimental design
• Identify pilot test against sampling of data
• Data quality measures and profiling process
• Data extraction or capture and transformation processes and responsibilities.
• Design the candidate visualization presentation and interactive discovery
• Set priorities and candidate desired timelines.

Acting

• Pilot the insight science discovery.
• Learn from the pilot and refine the plans.
• Implement the refined improvement.

Learning

• Monitor and control the data quality and visualization from the insight engine.
• Diagnose anomalies and adjust the knowledge representation.
• Improvement recommendation and iterate to gain further insights.

IDEAL presumes initial candidate solution from its inception based on the clients tacit knowledge representation and access to data that models the real-world activities. It works from the strengths of the assets the client has amassed through its existing data capture.  In this way it is different than other process improvement programs that focus on correcting weaknesses and removing cost. The framework is designed for incremental transitional improvement through opportunity discovery thus it is well suited for new business value creation through the application of scientific computation.

Data science, machine learning, pattern recognition, statistical processing, large scale computation against terabyte databases, these are all complex to implement.  We are experts in using these techniques to support the scientific research that our team is active in conducting. Opportunity seeking using computation and data with our Insight Science methods is one of discovery, monitoring the data quality against the knowledge representation and our clients gaining deeper understanding of their business through our visualization.  The client’s business domain experts can deploy dramatic new value with our team of scientist unleashing inherent value that simply is not possible without the computation and mathematical models we can develop for them.