VIA Analytics Data Flow

Single, Interactive Visual Modeling Paradigm for all IoT Analytic


VIA’s Analytic Models are built using the powerful Analytics Data Flow (ADF) environment that empowers developers and business analysts to rapidly create analytics-based solutions using visual models requiring little or no coding. ADF has a single visual modeling paradigm for streaming and batch applications over machine learning, descriptive, predictive, and prescriptive analytics. This visual modeling environment enables the rapid creation of IoT Analytics solutions in days, not months.


The Analytics Flow Designer within ADF provides powerful modeling capabilities based on the core concept of “analytic pipelines” consisting of multiple steps of analytic processing, where each step yields increasing analytic insight. A visual dataflow language enables solution developers to rapidly lay out an analytic pipeline, consisting of multiple data and analytic processing steps, with high level building blocks. The Designer accelerates solution development via a rich “drag and drop” library of reusable building blocks. Developers can also compose their own custom building blocks and incorporate them into the library of available blocks.

The built-in library of building blocks includes:

  • Data Sources & Target Connectors supporting protocols and data formats for a wide variety of data
  • Data Preparation, e.g., Filter, Parse, Transform, Enrich
  • Descriptive Analytics, e.g., Correlation, Statistical Summaries, Multi-dimensional Analysis, KPI Computation, Pattern Matching, Trending
  • Machine Learning, supporting a wide variety of regression, classification and clustering algorithms
  • Predictive and Prescriptive Analytics, based on ML models and supporting streaming (real-time), online, and batch processing
  • SQL Queries
  • SDK for encapsulating custom-built or imported code as an ADF block


Analytics Flows designed in the Designer are processed in the ADF Analytics Flow Engine — a run-time framework to manage Analytic Flow processing. Consistent with the visual modeling environment, the Analytics Flow Engine processes streaming and batch applications over descriptive, predictive, and prescriptive analytics. The Engine includes key features such as:

  • Event Time processing with support for processing out-of-order data and other timing anomalies
  • Time Series processing to provide insight into the behavior of IoT networks over time
  • Fast Geospatial Correlation
  • An Interactive testing and debugging environment
  • Seamless integration with the VIA Open IoT Data Lake
  • Full lifecycle management of the Analytic Data Flows


VIA’s ADF Environment not only streamlines the creation of analytics via its visual development environment, but also delivers elastic and scalable IoT analytics using a scale-out architecture. ADF takes advantage of VIA’s Foundation to provide elastic scalability, end-to-end reliability, and holistic lifecycle management. ADF delivers production-grade analytics processing for the most demanding IoT use cases.