[Talk-GB] Q4 2020 Quarterly Project: Defibrillators

[Cj Malone]

> Firstly, I have seen a few buildings that have an AED pictogram sign
> outside, suggesting that there is a defibrillator inside. Is this
> considered sufficient 'on the ground' evidence to add to the map.
> These are often locations that are noted up on the Survey Me! tool,
> but not always.

I would argue so, we add phone numbers to hotels from outside signs.

There is a question of if private defibs should be added to OSM, and
again I would argue for there addition. We aren't using these defibs
for routing in emergencies. We have no access liability with Ambulance
Services. These are added to OSM for education purposes, so that local
people can be better informed about there situation and so people can
process the data to workout areas that are under served, improve
coverage and save lives.

> Secondly, I notice that Rob's otherwise excellent Survey Me! tool
> occasionally incorrectly matches a point quite far away, and so flags
> up a missing defibrillator, even though it is correctly mapped in the
> location expected by the tool. Is there an easy way to resolve these,
> or is this just too complex a problem?

I don't think Rob has manual matching on his defib site, but I may be
wrong.

The underlying issue is the quality of the source data, they don't
typically have coordinates, just a postcode so the location isn't that
accurate. They also don't have refs so we can't match defibs.

Hopefully there will eventually be a central list of defibs that OSM
can work with and improve, BHF has hopes of doing this but it seems to
be a bit stagnant. I also hope this will be released Open Data, OSMUK
would be happy to help with that.

> Either way, it just highlights another reason why this too cannot be
used to add data to OSM.

The copyright situation is the external reason, it'll put OSMF in a bad
position. Copying this data is just as bad as using Google Maps as a
source.

> However, next time I am passing by the station, I think there is one
> missing that I can add. This might fix this incorrect matching…

I have noticed that the more defibs in OSM, the better the matching
gets. But that's not always the case.