Big Data Consulting

The existing client application is a cognitive cloud based platform designed to:

• Be used by researchers and clinicians along with CxOs to mine medical records and stay on top of new publications by converting unstructured real world data into actionable insights.
• Extracting high-value knowledge for further downstream research or treatment plans.
• Connect to hundreds of multiple data sources
• Build and publish complex data analysis pipelines
• Visually produce insights from complex data.

The end users however faced hurdles in:

• Slow computation of data
• Extraction of large volumes of unstructured clinical data
• Delay in job status tracking in visual pipeline manager
• Steep & continuous rise in data size
• Legacy system unable to produce quick results to meet specific clinical and research requirements

Team CSS decided to conduct a detailed analysis that addresses these challenges. In the first phase of the solution design after deep analysis of technology, business and process we identified the need to do a in depth analysis. Following were the major activities conducted:

• Technical evaluation of the legacy application
• Conducting a fit-gap analysis
• Identifying scalable components in the system
• Creating a visionary product roadmap
• Identifying the need to re-architect the existing solution
• Defining components that require replacement with faster and scalable computation capabilities

Post successful evaluation, discussion and brainstorming sessions, CSS provided a comprehensive analytical report. Following were the major outcomes of the new solution design:

• The big data architecture provides vertical scalability to extract large volumes of clinical unstructured data
• No horizons for increase in data size
• Complete real time visibility of data leading to quick actionable insights to meet specific clinical and research requirements
• The new architecture creates a roadmap for further downstreaming research and treatment plans
• Maintaining computation speed irrespective of size of data helping researchers to stay on top of new publications and medical records.
• Fast extraction of large volumes of data through multiple sources thus converting  unstructured real world data into actionable insights.
• Design of required hardware to support the big data eco system