Ambient data management systems have their beginning in the 1970s, with the introduction of minicomputers and small microprocessor-based data loggers. The previous generation of paper strip charts were whisked away or relegated to a secondary backup role, while the industry enjoyed the ability to acquire their data every day over 300 baud modems.

Microprocessors gained power, the data loggers became more powerful and sported local LCD interfaces, and the inexpensive DOS-based (and later Windows-based) PC replaced the expensive minicomputers and small mainframes, with the last ones just being replaced this year (2006). We are enjoying the front of a wave of less expensive and high bandwidth communications to our sites.

However, the industry remains stuck because every data system- every commercial ambient data management system on the market today, still depends on data loggers (including PC-based loggers) to act as the front end to the central data nexus.

AirVision will be the first system to eliminate this barrier and bring the industry into a revolutionary new way of managing data.


Expanding Data Handling Capabilities

We have moved into serial-based data acquisition, taking advantage of the power of microprocessors within the instruments themselves, but all vendors still depend on a data logger (again, including PC-based loggers) to do the front-end work. AirVision will be the first system to break the architectural barrier and treat every point on the network as a normalized “data source.”

Data collection should be just that- data collection, without any other representation in the application. Whether we are collecting averages from a data logger, readings directly from a serial/Ethernet capable instrument, collecting and processing data from a PM sampler (BAM, TEOM), or importing files from lab data, the end result can be managed independently from AirVision’s other functions. AirVision supports an open system of modular ‘drivers’ that can be added to provide connectivity to any source of data. The driver manages the details of data collection and uses standard interface to exchange data with AirVision’s core. Third parties and end users can construct new drivers for AirVision, providing an open solution to manage future requirements. AirVision is being designed specifically along the concept that eventually networks (or significant parts thereof) will consist of smart instruments connected to a central AirVision data management hub through broadband connections [see Agilaire’s presentation, 2006 Monitoring Conference].

Air Vision System Architecture

Powerful and Expandable Data Post-Processing Capabilities

Existing systems provide little post-processing of the data, usually limited to data editing and other manual functions. As the quantity of data grows and staffing pressures increase, agencies and consultants need to optimize the quality assurance process as much as possible.

AirVision opens up the process between data collection and final reporting through an open modular approach similar to Open Data Sources. Open Data Processors can be scheduled and triggered as defined by the user in the Task Manager to automate quality assurance while also controlling the points of data access and display, such as AirNow and web presentation (See diagram).


These post-processing tasks might include:

• AQI calculations
• AQI forecasts
• Zero or other calibration data adjustments
• Automated AQS coding
• Web publishing – tables, real-time charts, etc
• AirNow or real-time Exchange Network transfers
• E-mail alarms
• E-mailed reports
• Time skewing tape sampler data

The open approach also ensures that even known data paths can be easily updated to meet changing requirements, such as future alterations in the AQI calculation process (changes based on new standards, and perhaps, we envision, more complex algorithms that take into consideration the synergistic health effects of PM and ozone, requiring a multivariate algorithm to calculate a more representative measurement of health effect).

Automated Quality Assurance

Key to this strategy will be the Automatic Data Validation Processor Module. The ADVP assigns a quality grade (1-10) for each data point based initially on instrument / data logger flags, but allows the user to generate any number of “rules” that affect the quality code (Diagram shows examples). The ADVP can be triggered after data collection to grade each data point, and the quality grade can be used as a gate to prevent suspect or bad data from being published to the web, shared with other organizations, etc. The grades can also be used to focus quality assurance efforts on the most suspect data points.



Data can be compared against other parameters at the same site, different sites (spatial testing), or even against historical composite values for that particular parameter and/or site (e.g., “how does this hour’s value compare to the same hour, this week, over the previous five years?”). Persistency checks are also included as part of the ADVP library of functions.

The ADVP can be layered with other Open Data Processing modules (such as cal zero/span correction) and reporting/presentation tasks (AirNow, Exchange Network XML transfers) to provide a well controlled and documented process of data flow.


Consideration of New and More Complex Data Relationships

In numerous customer interviews, it became apparent that, among other complex data relationships, the need exists and is growing for managing data from parameters in a hierarchical fashion. Consider the case of PM filter data that may exist both as a total mass/concentration value, but also may consist in detailed breakdowns from speciation monitors or XRF lab data. The speciated values should not exist as simply separate database parameters, but should ideally be tied to the overall mass/concentration value in a “parent-child” relationship. AirVision supports these relationships as part of its original design, a feature unique in its industry, and not viable as an “afterthought” or modification to an existing database.