For the past year the Norwegian Allergy and Asthma Association (NAAF) has been working on upgrading its pollen count warning service. It has benefited from a much needed facelift, and the system has been entirely rebuilt with modern platform technology on a scalable cloud solution. Itera is behind the solution and is now looking at how existing pollen data and machine learning can be used to further improve pollen count warnings. 

Itera naaf pollenvarsel 1920x1080

modernised solution focused on the user 

Our pollen count warning service has existed for a long time and had looked the same for quite some time. We needed a new solution that would take into account the requirements of universal design, innovation and user friendliness, explains Hogne Skogesal, Director of Communications and Marketing at NAAF. 

To start, Itera analysed NAAF’s existing solution and advised on how the new solution could be built. End users now have roughly the same functionality as before. They can check pollen levels on NAAF’s website or choose to automatically receive a warning when high levels of the sort of pollen to which they are allergic are likely to be present in their area.  

The pollen data is used in several different ways. A range of organisations currently subscribe to a data stream from the service, which enables media companies to provide pollen warnings as part of their weather and news services. Other organisations use the warnings for analytical purposes. Pharmacy chains can, for example, use the data to plan their logistics and how much pollen medicine to have in their warehouses. The new solution uses the same format for the data stream as before, meaning data customers did not need to change their systems. 

NAAF og Itera (1).jpg

Hogne Skogesal (Director of Communications and Marketing at NAAF), Ingrid Rise Fry (Communications Advisor at NAAF) and Martin Smestad Hansen (Senior Consultant at Itera). 

 

Machine learning and historic pollen data  

We have looked at how our pollen count warning service could be improved and have thoroughly evaluated it. From the resulting report, it is clear that the service could be significantly improved”, comments Hogne Skogesal. 

There are several possible areas for improvement, including the channels that are used to communicate the warnings and the warnings themselves. Itera’s own specialist machine learning and AI team is now working on analysing historic pollen data and data from other sources to look at the options in terms of making the warnings more precise and longer-range.  

 

The way ahead: improving the measurement process 

The most challenging aspect of the warning process at present is measuring pollen levels. Norway has too few measuring stations, and the process is also very manual and therefore not easy to scale up. 

The underlying technology simply needs updating. Currently a physical sample has to be gathered from the traps, and this is then posted to our pollen researchers. Digitalising this is key to improving the process going forward, and it is exciting to be looking at what solutions are available as part of our project with Itera, concludes Hogne Skogesal. 

There are often big local differences in pollen levels, and the existing 12 stations are therefore insufficient. Itera is also currently working on investigating suppliers of entirely automated measuring stations and is also assessing the options in terms of being at the cutting edge by commissioning the development of cheaper and smaller measuring devices. This is one possible way of increasing the number of measurement points and thus making the measurements used more precise.