Implement statistics API

[nuest]

The API was designed from the client perspective and is documented in an issue for enviroCar-www, see enviroCar/enviroCar-www#109

API sketch transferred from the enviroCar-www issue is below, let's continue discussion here.

• /api/statistics/
  • already exists
  • change into a list of available statistics resource (like https://envirocar.org/api/stable/) and provide the content under api/statistics/phenomenon/:phenomenonid
• /rest/statistics/project
  • returns the overall statistics listed above, example below
  • should use LOD identifiers if possible, for example http://sweet.jpl.nasa.gov/2.3/reprSciUnits.owl#kilometer .
• /api/statistics/project?fields={:id}
  • return just the information for one identifier (value, id, description, unit)
• /api/statistics/project/:year/:month/:week?fields={:id}
  • the same information as above but only for specific months or days
  • examples
    • /rest/statistics/project/
    • /rest/statistics/project/2013/8/23
    • /rest/statistics/project/2013/8?fields={emission:co2,tracklength}

{
   "source":"www.envirocar.org/api/statistics",
   "statistics":[
      {
         "value":12345.6,
         "id":"tracklength",
         "description":"The accumulated number of the kilometers of all tracks on the enviroCar server.",
         "unit":"km"
      }
   ]
}

[nuest]

architecture on the server side:

• statistics information must be stored somewhere and cannot be calculated on demand for all requests
• statistics could be
  • updated with every track upload
  • updated with a nightly task
  • updated manually

[nuest]

Regional statistics, e.g. for a local community, a country, continent or all data, could be encapsulated as WPS processes and should be cached. This would also allow a "divide and conquer" approach to statistics: Calculating stats for Europe results in calculating the stats for each country, (for each Bundesland). All cached statistics are only updated if a new track was added (comparing the hashes of a list of track ids?).