Hi Craig, don't worry :) i can wait. I'm thankful for every input i can get on this topic.
The way you explained peter's solution brought it closer to me. It could really work out for my use case but i have to test it. My concerns are: is neo4j strong enough to handle - lets say for example - 50.000 cars which are uploading their positions and some extra data every 5 till 10 seconds and manage the extra nodes containing the amount of cars on the related way? So the data of my database is except for the osm base data very volatile and changes occur really often (cars are static, but their positions in my graph are changing as they are driving on). I want to inform the cars as fast as possible when i detect a traffic jam on my database, so the application has to work in nearly realtime. Is this possible that way? Thanks again for every idea and solution :) you are doing a great job guys! Chris On Wed, Apr 27, 2011 at 9:15 PM, Craig Taverner <cr...@amanzi.com> wrote: > Hi, > > Sorry for a slow response, I've been travelling. > > I think I'm missing a key point here. I am not aware that the OSM model > includes the concept of the number of cars at a point, or on a way. So this > means, IMHO, that you need to add this information to the model. So, if, > for > example, you have a table of data somewhere that lists street addresses and > how many cars that street has (no idea where you get this data, but > obviously you need it to do your analysis); then when you add that data to > the graph, make sure you add it in a way that will support your query. > > Peters idea is just such a suggestion. Since the number of cars, or density > of traffic, or whatever property to add, could simply be added as a tag to > the way, you could also index that property in a lucene index, and query > the > index later, for all ways with traffic > some threshold. Another option > would be to link the tags (or ways) for which the threshold is passed to > some new node you made, so you can just query that nodes children for the > results. Expanding this idea to instead make a set of new nodes, each > representing a range of traffic values, and link all the ways to the node > representing the traffic range they are in, and you suddenly have a > histogram or bar-chart of the distribution of traffic over the ways. > > Did I understand this correctly? > > Regards, Craig > > On Tue, Apr 26, 2011 at 7:58 PM, Christoph K. < > klaassen.christ...@googlemail.com> wrote: > > > Hi Peter, > > > > i think your approach does not exactly match my use case. To be a bit > more > > precise: i want to determine a traffic jam. Therefor i thought it would > be > > great to have a mechanism like "give me all regions where the count of > cars > > is greater than x" and then i do further inspections of every region to > > determine if there really is a traffic jam. > > A not so nice option would be your counter. But as far as i understand > your > > suggestion, i have to put a counter property on every way of my osm map > (in > > my example Bavaria -> Germany) and have to manage this counter. This - i > > think - would be pretty much workload for the server, as there are pretty > > much cars and ways at a single point of time. > > > > Would be nice to hear further thoughts of yours on this topic :) > > > > Greetings > > Christoph > > > > On Tue, Apr 26, 2011 at 6:37 PM, Peter Neubauer < > > peter.neuba...@neotechnology.com> wrote: > > > > > Christoph, > > > you could just update some counter and index on that when you are > > > indexing an entity into a particular bounding box or other area of > > > your map? Do you have any definition of what a "spot" is? Is that a > > > fixed area? In that case, this approach might work? This would be an > > > approach of updating this index when things are moved/inserted into > > > your structure, but still I think that might work out... > > > > > > Cheers, > > > > > > /peter neubauer > > > > > > GTalk: neubauer.peter > > > Skype peter.neubauer > > > Phone +46 704 106975 > > > LinkedIn http://www.linkedin.com/in/neubauer > > > Twitter http://twitter.com/peterneubauer > > > > > > http://www.neo4j.org - Your high performance graph > > database. > > > http://startupbootcamp.org/ - Öresund - Innovation happens HERE. > > > http://www.thoughtmade.com - Scandinavia's coolest Bring-a-Thing > party. > > > > > > > > > > > > On Tue, Apr 26, 2011 at 9:45 AM, Christoph K. > > > <klaassen.christ...@googlemail.com> wrote: > > > > Hey guys, > > > > > > > > i'm back on the track and try to do some things with neo4j Spatial. > > > > As the subject of this post tells, i'm looking for a way to find a > spot > > > on > > > > my OSM map, where more (or less, or equal) than x > cars/people/buildings > > > are > > > > located. > > > > Is there a way to perform this task without traversing my whole map > > > > rectangle by rectangle? > > > > > > > > Use case: i try to find bottlenecks dynamically on my osm map which > is > > > > determined by an amount x of cars in a region y. > > > > > > > > Thanks for your help =) > > > > > > > > Greetings from Munich > > > > Christoph > > > > _______________________________________________ > > > > Neo4j mailing list > > > > User@lists.neo4j.org > > > > https://lists.neo4j.org/mailman/listinfo/user > > > > > > > _______________________________________________ > > > Neo4j mailing list > > > User@lists.neo4j.org > > > https://lists.neo4j.org/mailman/listinfo/user > > > > > _______________________________________________ > > Neo4j mailing list > > User@lists.neo4j.org > > https://lists.neo4j.org/mailman/listinfo/user > > > _______________________________________________ > Neo4j mailing list > User@lists.neo4j.org > https://lists.neo4j.org/mailman/listinfo/user > _______________________________________________ Neo4j mailing list User@lists.neo4j.org https://lists.neo4j.org/mailman/listinfo/user
