Pollen count answer provides information on the current pollen count, top plant allergens and additional pollen and allergy specifics to the user at the requested time and and location.
Research conducted prior to the design process revealed that United States (US) has the highest number of pollen related searches and other English speaking countries do not seem to be searching for this information as frequently. It was identified that top pollen search hot spots are Texas, Atlanta, Ohio, New York and Seattle with number of searches rapidly increasing in spring and early summer months and reducing later in the summer.
Our first set of explorations focused on using similar approach to other pollen tracking websites. We used a lot of circles and “traffic light” color system, where green color meant low pollen levels and red meant that pollen levels are high. With a forecast being required it made sense to use similar styling to existing Bing weather answer, which meant more circles. We soon realized design became too busy and complicated to read, so we started looking at swapping circles into various icons. We also tried using combinations of circles and bars, but use of different elements in order to show the same information did not seem to make a lot of sense either.
During the very first stages of design process we agreed that use of so many colors does not reduce the amounts complexity and is definitely not helpful for people with color vision deficiency. As there was no previous design in place, we felt like we could go a bit further and do something more exciting.
After more iterations on icons and colors there were several other concerns raised and one of them was weather we should or should not include numerical values for pollen severity. From previous research we knew that a lot of other pollen tracking websites relied on color as a main severity indicator, but we found out that the ones that use numerical values had completely different scales. By creating a survey we wanted to find out if adding numerical values would help and inventing our own system would not confuse potential users.
Total of 100 respondents participated in our survey and there was only 2% difference between two groups. Although some users found numerical values useful, most of them were getting confused by a non standard range of 0-4.