AMA data about Milan
Urban network, traffic data and O/D matrices by AMA (Agenzia Milanese Mobilità Ambiente - Milan Mobility and Environment Agency)
AMA ontology and data
OWL ontology (includes RDFS comments with useful explanation of the concepts and proprty modelled)
RDF dump (on SVN, requires password)
Additional notes about some concepts and terms of the ontology
FAQs
Q: What is the meaning of BRV curve?
A: It is the Bureau of Public Road Function. You can find more information here
Q: The usage of "Prohibition"; Should we assume that the only explicit route under "Prohibition" is all prohibition route or should we assume that the information is only known information for prohibition? (CWA vs OWA)
- A: That's a good question. Actually we don't know the exact answer: the prohibitions you refer to are all the information we got from AMA, but we don't know if they are complete. We think that for now we can assume a CWA if it is easier for you to process those data.
Q: The distinction between lFrom and oFrom still are not clear to me. If tFrom under Traffic class has another meaning compared to lFrom and oFrom, can you compare those 3 properties?
A: The three *From properties has different domains/ranges. You can find more information about them here
Q: Regarding to ODrecord class, how to get reason? In other way, how to get ODrecord instance? My understanding is that ODrecord class represents vehicle movements with time and reason. I can not imagine how to generate those data unless all vehicles have each own sensors and input from drivers.
- An OD matrix indicates the number of vehicles that move from a zone to another one. In addition AMA added a set of parameters like the reason of the movement and time information. As you said, sensors in cars are one of the possible ways to get this data, but AMA realized OD matrices by interview: in 2005/06 they contacted about 10% of people living in Milan or in adjacent towns and asked them about their travel.
Q: overflowSpeed property under Link class is dynamically changeable? Does it represent the speed on the specific time or average time? Also, I wonder the units of values.
- A: overflowSpeed is expressed in km/h. It is fixed and it is a feature of the street, it is the outflow speed when the street is free (in other words when it is not congested)
Q: Is this ontology a whole part or a part of the whole ontology?
- A: The ontology represents the part of data that we actually converted and exposed in RDF.
Q: Could you explain more about tEquiv comparing with overflowSpeed property under Link class?
- A: The definition of tEquiv, which is a property of Traffic class, is: "Vehicular flow values measured by equivalent vehicles passed between 7:30am and 8:30am of a medium weekday" and it is expressed in number of equivalent vehicles, so it could not be compared with outflowSpeed, that is a speed (expressed in km/h). You could compare tEquiv with property capacity of Link class, that is expressed in the same way (number of equivalent vehicles). The difference between these two properties is that capacity is a static value and it depends from the feature of the street; tEquiv is a statistical value obtained from measurements done by AMA, and it is a value that could change over time.
Q: How to recognize "Jurisdiction" with coordinate of node?
- A: Jurisdictions are described as "sets" of links and the links are delimited by nodes; therefore some nodes will be part of a single jurisdiction, while other nodes will be somehow "on the border". Anyway, we also have other information about "zones", which are expressed by polygons of nodes, but we don't have yet the respective RDF version of those data. We will keep you updated on this point.
Q: How to express situations (context information like weather) with this AMA ontology? Do we have to add?
- A: Yes, for every additional piece of context information, the ontology must be extended or enriched. The current version of the ontology covers only the data we currently translated into RDF.
Q: How to connect with a map with each node? Is it already developed with other mechanism?
- A: For each node, the data contains its coordinates (latitude and longitude) that can be used to connect the data to a map
Q: Can we get coordinates value from the linked data? We'd like to know a directions(left, right, south...etc)
A: Each node has its coordinates. The direction is not simply north/south or east/west, because Milano topology is much more complex (see e.g. http://tinyurl.com/67ond2 to see Cefriel neighborhoods).
SPARQL endpoint
SPARQL endpoint (requires password)
Example queries
- Show the triples of links that starts from node 14099 and that arrives to node 14101
- SELECT ?l1 ?l2 ?l3
- WHERE{
?l1 lud:lFrom <http://seip.cefriel.it/ama/resource/nodes/node14099>.
- ?l1 lud:lTo ?x.
- ?l2 lud:lFrom ?x.
- ?l2 lud:lTo ?y.
- ?l3 lud:lFrom ?y.
?l3 lud:lTo <http://seip.cefriel.it/ama/resource/nodes/node14101>.
- }
- Show all reachable nodes from node 1517 with 1 or 2 links
- SELECT ?n
- WHERE{
- {
?l1 lud:lFrom <http://seip.cefriel.it/ama/resource/nodes/node1517>.
- ?l1 lud:lTo ?n.
- }
- UNION
- {
?l1 lud:lFrom <http://seip.cefriel.it/ama/resource/nodes/node1517>.
- ?l2 lud:lFrom ?x.
- ?l1 lud:lTo ?x.
- ?l2 lud:lTo ?n.
- }
- {
- }
- Show the couples of nodes that have 3 or more years of traffic relevations
- SELECT DISTINCT ?f ?t
- WHERE{
- ?n1 lud:tFrom ?f.
- ?n1 lud:tTo ?t.
- ?n1 lud:year ?y1.
- ?n2 lud:tFrom ?f.
- ?n2 lud:tTo ?t.
- ?n2 lud:year ?y2.
- ?n3 lud:tFrom ?f.
- ?n3 lud:tTo ?t.
- ?n3 lud:year ?y3.
FILTER(?y1 < ?y2).
FILTER(?y2 < ?y3)
- }
- Count the number of links for each link type (highway, main road, secondary road, local road)
- SELECT ?x COUNT(?l)
- WHERE {
- ?l lud:linkType ?x.
- }
- GROUP BY ?x
- Count the number of O/D records for each couple time interval-reason
- SELECT ?w ?r count(?od)
- WHERE{
- ?od lud:when ?w.
- ?od lud:reason ?r.
- }
- GROUP BY ?w ?r
- ORDER BY ?w ?r