SBIR Phase I: Scalable Collaborative Analytical Modeling

Period of Performance: 07/01/2017 - 06/30/2018


Phase 1 SBIR

Recipient Firm

Fact Labs Inc.
1864 15th St Unit 204 Array
San Francisco, CA 94103
Firm POC, Principal Investigator


The broader impact/commercial potential of this Small Business Innovation Research Phase I project is to enable organizations - whether businesses, governments, or non-profits - to make more informed, more data-driven decisions. All organizations must decide how to allocate limited resources and do so in the context of meeting a set of objectives, such as profit, social wellbeing, or health. Modeling as a process and models as artifacts of that process allow decision makers to understand data through the lens of objectives to then make decisions; data alone, no matter how much, cannot make decisions. As more aspects of the world are instrumented and captured digitally, the breadth and quantity of data will out of necessity require more modeling to be codified and bring more stakeholders into the fold. Organizations need a modeling workflow and supporting tools that are capable of handling this wider range of data, are fully accessible to non-technical users, and allow more stakeholders to participate in this important process. This Small Business Innovation Research Phase I project addresses the challenge of many users collaboratively building and maintaining analytical models that are consistent and reproducible while allowing for divergent and convergent change. On one end, spreadsheets serve as a general-purpose, ad hoc modeling tool that is open-ended and accessible for many, and on the other, whole software applications whether packaged or custom developed are generally more powerful in important ways but usually sacrifice accessibility and generalizability. This project will produce a prototype of a collaborative integrated development environment for modeling that manages code and data. The technical feasibility of this prototype will be evaluated by developing a test framework that simulates the divergence and convergence of models and scores the outcomes. The commercial feasibility will be evaluated by testing with users building real models to understand the scope of functionality required to bring this to market.