[OSM-dev] real-time traffic map
bogdan.bivolaru at gmail.com
Thu Aug 21 07:39:10 BST 2008
I knew there will be privacy concerns, but we can try to anonimize data
* we can use personae, request a new persona identity every 1-2 weeks
* we send real location data over some sort of UDP over Tor over mobile
internet - I think this secures customers against malevolent OSM admins
* if we also use some sort of PKI encryption associated with the persona we
can also protect clients from carriers and crackers snooping over customer
In the end I agree this depends heavily over the size of our userbase, but
maybe we can increase our userbase with other services:
what if I wanted to send a friend a live 'feed' of my current location?
Wouldn't it be a great tool for friends trying to find each other? I thought
of doing this over the Jabber network and I think it works well for 1-2
users. Larger friend networks need sort of a server that aggregates the
location of several users and then sends location updates + maps in packs.
Of course there is one _small_ problem: I don't know how to protect the
privacy of the users in this situation. What if you want to talk with
someone and be invisible for others?
Larger picture in my mind:
I think there could be a lot of similar location-related services that could
increase our userbase and could interest money-wise entities around the
world: what if we could make a map that displays price cuts in shops in an
area? Or 'find me the closest shop that is still open for another hour at
23.44 local time'? Of particular interest I find that OSM can deliver this
kind of services worldwide.
So how should we deliver on these ideas? How do we make real, usable, cheap
services from ideas related to location and OSM? My particular worry is that
proprietary service providers (closed maps, proprietary software) will
deliver this sort of instantaineous services and OSM will become just "a
project that could be..." and everyone will forget about it.
In other words: I think this is a big opportunity for us, one that we should
not miss. What do you think?
On Thu, Aug 21, 2008 at 8:32 AM, Adrian Vogt <Adrian at vogt.net> wrote:
> Dear Bogdan,
> I loosely remember a short report in a scientific TV show (Menschen
> Technik Wissenschaft on swiss TV) reporting that Swisscom has provided
> *anonymized* cell-phone data to students for an evaluation on whether
> that data could be useful in traffic analysis. As far as I remember
> that report, the where successfully in interpreting that data and
> everything looked VERY nice. The only problem is that at that time,
> Swisscom did not see a real benefit in operating an infrastructure,
> which delivers this kind of data continuously.
> But I still believe this would be the most cost effective way to get
> such data, as there are many (millions) of subscribers and there is no
> cost for those providing the data (the people "running" around with
> their mobile phone). The investment of mobile phone company could be
> compensated by adequate pricing of the data (or making it free as a
> marketing instrument). Depending on the exact architecture, the time
> resolution could be quite fine (e.g. 15 minute intervals or even
> small). As there are so many data-points available by using the cell-
> phone network data (of course in anonymized form), you can even detect/
> separate trains (300 cell phones moving in sync) from motor-ways
> (thousands of cell phones in constant flow).
> I am not 100% sure, but their thesis should be linked as:
> . Unfortunately its written in German, so you will have to run it
> through a translation services if that is an issue.
> On 20.Aug.2008, at 12:00, Bogdan Bivolaru wrote:
> > Dear all,
> > Do you know any way to maintain an instantaneous map of a road? By
> > that I mean: can all the drivers using OpenStreetMap + GPS upload
> > their current location and direction they are going to, instant &
> > average speed to some server and then can the server process the
> > uploads to show the average speed and traffic?
> dev mailing list
> dev at openstreetmap.org
"The best way to predict the future is to invent it.", 1971, Alan Kay:
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the dev