It's interesting viewing, especially as it's live self reported data. I've even been able to use it to watch the snow move up the country to us in Birmingham. However, there are some issues I think.
- The default 'snowflake' icon view is a bit misleading, I suggest setting to the red or blue large snowflake, which makes it easier to see the size and frequency of tweets. It's found in the Settings in the top right.
- The map isn't a map of snowfall in the UK, but a map of people tweeting about snowfall in the UK. Given he central square mile of Birmingham will have more tweeters than a central square mile of the Yorkshire moors, of course there will be more tweets in built up areas. And more tweets equals more snowflakes on the map, making it look snowier than it is, when in fact those multiple snowflakes are just multiple reports of the same snowfall.??
- However, given the above, on the widest view, showing the whole UK, it still seems to give a pretty accurate summary of the current snowfall.??
- Finally, I have issue with the 'self-reportedness' of the project. Of course the nature pof the project is to allow people to report on what they see, but it's all very subjective. People have been reporting 10/10 snowfall, which means it's snowing the heaviest it could. I'm sure if that same person then went to Alaska or somewhere very snowy, they might revise their UK 10/10 down to 8 or 9. I'm sure even heavy snow we see here isn't that heavy.??It would be interesting to position a handfull of people at a window view of snowfall and ask them to rate what they see out of 10 – I wonder what kind of variance we'd see???
- One convoluted (and overkill probably) solution would be the addition of a photo feature. The person reporting their local snow takes a photo of snowfall against a large sheet of black card held at arms length, the idea being that this shows the size, frequency and amount of snow. The image is then uploaded, and some clever image-reading algorithm is then able to rate the snowfall on a mathematical, data-driven basis – it's not subjective. Basically, answering the question, how much white can I see against this black? As before, a x/10 rating is then applied to the map. ??