Skip to content

NovaStar Data / Data Quality


Introduction

Ensuring real-time data quality can be challenging. Legacy ALERT flood warning systems are particularly impacted by data quality issues because the transmission protocol can result in data collisions and corrupt data packets. ALERT2 and other data collection technologies that minimize data corruption result in higher data quality.

Data Quality in NovaStar

The NovaStar system attempts to filter out bad data by comparing data reports against range limits and examining sequential data reports. Data values that are clearly out of range are marked as such. However, identifying bad data in cases where a storm or some other rapid change is occurring is more difficult and sometimes bad values do slip through. Each data report has a data flag that can be retrieved with data web services and used for analysis and display. The Q flag indicates a questionable value and V indicates valid values. The NovaStar system provides features to edit bad data and recompute derived data values so data values may change over time.

Data users should also be aware of the inherent errors in data collection. In particular, river discharge and other values that must be estimated have higher error. Each system and each station's data quality (accuracy) will vary depending on equipment, maintenance, and local conditions. For this reason, data users should understand potential issues with using data for decisions such as water supply allocation. It is generally best to evaluate data from multiple stations and sources when making critical decisions.

Correcting Data Quality Issues

Bad data can be corrected using NovaStar Administrator web application or other applications. For example, the TSTool software can be used to detect bad data values.

Services

The following services are related to data quality:

  • dataFlags - data flags used in data records
  • data - return data reports with flag
  • ts - return time series with flag