Collaborative Analytics
Predictive Analysis Without Integrating Data
Optimize Business Performance from Distributed Big Data
The “Big” Problem
The most significant barrier to realizing the value of combined big data from high variety sources such as mobile devices, social networks, and sensors is the need to integrate the data to perform analysis. Data integration for predictive modeling is complex, expensive, and challenged to keep pace with data dynamics. In addition, communicating and assembling large integrated datasets for machine learning is increasingly in tension with proprietary and security considerations. What is needed is a cost-effective means of quantifying the value of information and integrating diverse analytics into a unified model, without ever integrating the data.
Who Has It
All businesses are driven to use data to improve customer experiences through tailored or personalized content, to enhance products and services, and to improve efficiencies and lower costs. Over two-thirds of organizations are trying to blend together from five to fifteen different sources of data for analysis per a 2015 HBR report. Maximizing ROI on data analytics projects means targeting the spending to deliver answers at the required fidelity as cheaply and quickly as possible. Increasingly, small and medium enterprises are realizing the business opportunities locked up in their data, e.g., for targeted marketing or compliance risk monitoring, but they may be challenged to understand what they can do with the data they have, and what additional data they need to be collecting to meet their goals. Within larger enterprises, it can be the case that large up-front investments in data platforms dig deep holes of cost that the analysis results are challenged to ever backfill and deliver return on. Business leaders may believe there will be value in bringing together all that data, but they don’t know how much value they can realize until they aggregate it and analyze it, with the attendant delays in time-to-results further impacting the value proposition. The analytics lifecycle can become high cost and low agility. And once in place, the next level of work and spending spins up to sustain that commitment; meanwhile, the world around them and all the data that speaks to it just keeps on changing, so follow-on costs to maintain and build upon previous investments remain painfully high and they don’t abate. A compounding factor is that because the desired raw data content itself is often proprietary, or has critical business value for competitive advantage, there can be significant barriers to sharing it in a B2B setting.
A New Approach
Prism's technology enables cross-silo predictive analysis from distributed data, even in cases where access is restricted because it is private or proprietary. The method is based on a novel system of Collaborative Analytics in which the data is analyzed globally but in distributed fashion, in place, organically and privately. This is accomplished through a unique analytics platform that designs itself to answer user queries, and can run asynchronously, at global scale, over the Internet.
The Value
Because the raw data are never integrated, substantial acceleration is achieved in model learning and prediction, leading to superior agility to adapt to dynamic data, and significant reductions in system cost and complexity. The approach makes it possible to explore rapidly many combinations of data to make better decisions, even when sources cannot be directly accessed due to privacy or proprietary restrictions. You garner the "power of reach" across widespread networked data sources. The focus is returned to the framing of business questions that the available data support answering. You make better decisions faster and cheaper. These benefits provide value across many organizations and industries.
Concept Illustration: the business problem is to assess risk on a commercial loan using distributed high variety data. The Collaborative Analytics approach is to learn a private predictive network model across analysis silo’s performing independent factors analysis, over the open internet and in automated fashion, without integrating data. The benefit is better assessments, faster and cheaper.
The Products
The Collaborative Analytics platform enables real-time predictive decision making, with application to risk analysis, event prediction/detection, and business intelligence functions such as customer targeting and churn estimation. The capability is embodied in enterprise software that provides distributed information services across data silos.
With Collaborative Analytics the currency of exchange is information, in the form of privacy-preserving compact analytics and models, and not data, clearing pathways to new business models and marketplaces for information that allow both data consumers and data producers to trade confidently and profitably in all varieties of information.
The Services
We can help you define an analytics strategy for distributed data, design information processing architectures to your business requirements, and operationalize system solutions that leverage the latest methods in machine learning and artificial intelligence.
Realize the Value in Your Data. Make the Analytics Pay.
Copyright Prism Informatix 2020