Starting a Computational Ecosystem from Scratch

Problem  

A client’s computing ecosystem was antiquated and unfit for purpose.  Research scientists were frustrated because data was siloed and required time-consuming manual data wrangling to perform meaningful cross-experimental data analysis.  The IT organization had spent millions of dollars on a multiyear IT solution that met no scientific needs.

Solution

We performed a comprehensive assessment of the client’s IT needs.  We engaged with senior research executives, scientific and computational thought leaders, and bench scientists to document scientific and data analysis workflows and pain points.  We jointly developed a prioritized list of strategic opportunities and generated a comprehensive program roadmap and business case. Funding proposals were successfully presented to Discovery Research, R&D, and CIO Leadership teams.

While promising transformational change, we needed time investment from already over-taxed scientists who were skeptical of success the second time.  We had to demonstrate quick success, generating credibility as we went.

Outcome

We uplifted source systems, improved data management practices, and introduced a scientific workbench that enabled self-service access for scientists to easily traverse the data point space to execute data analyses, then capture and share knowledge learned.

Overall, we:

• Created institutional memory.
• Improved data processing efficiency, data quality, and ease of data consumption.
• Enabled end-to-end traceability of data for experiments and protocols at all levels.
• Retired inflexible legacy application with a better system that adapted to rapidly evolving business processes with minimal IT or business support.