With current legislation and guidance placing increasing strain on management resources, the provision of monitoring data for compliance and performance reporting has never been so important.
The primary stresses on existing data management systems are the volume and complexity of data collected, the interpretation of data trends, and the compilation of monitoring reports. Essential components of a data management software package thus include ease of data entry, strict control of data quality, establishment of an audit trail, automatic exception reporting, and visualisation of trends.
The following are selected tips to promote better data management disciplines which should ultimately save you both valuable time and money.
Data entry
Not even the most powerful mainframe computer on earth can resolve data inconsistencies. The data manager is responsible for ensuring data integrity and enforcing rules on how to deal with ‘problem data’ such as missing values, faulty measurements, transcription errors etc.
- To reduce transcription errors, obtain your data in a compatible digital format, for example via e-mail or on floppy disk. Manual copying of data from paper to computer is time consuming and prone to error.
- Do not call sample point 1 ‘spt 1’ one day and ‘spt1’ the next.
- Accurately record the date and time of sampling and analysis.
- Keep determinand names the same, for example, call methane either ‘methane’ or ‘CH4’, but never both.
- Ensure that the same units are always used for the same determinand even across reports. Differing units can play havoc with any subsequent analysis!
- Don’t permit zeros to represent empty records. If you must have an empty record, ensure it is a null (blank).
- A proven way to ensure data consistency is to provide all of your data suppliers with templates to be filled in and returned.
Data storage
Data is often just stored in files where the format may be idiosyncratic. A better solution is to take advantage of a standard database management system. This helps with data structuring and allows more convenient manipulation.
Having all this data isn’t a bad thing – when properly managed, it is an invaluable source of information, highlighting real or potential problems before they occur. But managing it effectively is not as simple as just collecting it.
Managers often receive data from multiple sources on a weekly or even daily basis – field readings, labs, met stations, etc. They are then expected to check the data for breeches of consent and erroneous or rogue readings, and then collate it all to produce reports, graphs, trends, etc. This can all be very time consuming, laborious, expensive and maybe most important of all, potentially unreliable – attempting to join together a multitude of different spreadsheets every month is the sort of task that is prone to human error.
Single repository
Keep data in one place. Multiple files make analysis more difficult. Use spreadsheets wisely. Don’t start a new worksheet for each set of monitoring data. Databases are better, but if you must use a spreadsheet, keep data files as large as possible and extract required data using the query tool or pivot tables, etc. If you choose the database route, ensure that is well designed by a professional. Badly designed databases are normally better than spreadsheets for data storage, but well designed databases are faster, and ensure a higher level of data integrity.
As digital as possible
Automate as much of your data management as possible. Dedicated data management software is invaluable – it can save vast amounts of time and provide a more accurate output. There are very few off-the shelf systems available that are flexible enough to adapt to individual situations. Key qualities to look for are:
- Ease of use;
- Flexible data import;
- Flexible addition of determinands and units, thresholds and other system parameters;
- Capable of holding any type of monitoring data;
- < and > symbols should be catered for; and
- Reports should be flexible.
The latest data management package from Monitor-Pro, Monitor-LT, is highly adaptable, and available for the management of all monitoring data.