Time depend speed values for edge from geo coordinates

I am doing research about including time depend speed matrix into graphhopper and any possibilities to do it.
Case description:
In our team we are managing a set of cars driving over a city during working hours.
Now we collected positions data from more than one year, and based on that we would like to create our own time depended speed matrix.

After digging in this topic on forum I’ve found a post, which I think could solve our problem. But according to that I have got several questions:

Map-matching
After converting our raw location data into .gpx xml file we can use map-matching tool.

  1. How can we export parsed road and what is an inner file structure of that informations?
    I checked regular graphhopper with one road request and I’ve checked json structure of output, but there are no informations about way_id or any other information we could assign this specific segment to the graph i.e:
	"hints": {
		"visited_nodes.sum": 6,
		"visited_nodes.average": 6.0
	},
	"info": {
		"copyrights": [
			"GraphHopper",
			"OpenStreetMap contributors"
		],
		"took": 5
	},
	"paths": [
		{
			"distance": 27.724,
			"weight": 3.53481,
			"time": 3118,
			"transfers": 0,
			"points_encoded": true,
			"bbox": [
				18.669013,
				50.311236,
				18.669155,
				50.311468
			],
			"points": "smarHih}pBq_k@l@[?",
			"instructions": [
				{
					"distance": 27.724,
					"heading": 158.61,
					"sign": 0,
					"interval": [
						0,
						1
					],
					"text": "Kontynuuj na Toszecka, 901",
					"time": 3118,
					"street_name": "Toszecka, 901"
				},
				{
					"distance": 0.0,
					"sign": 4,
					"last_heading": 158.6061768481145,
					"interval": [
						1,
						1
					],
					"text": "Jesteś u celu!",
					"time": 0,
					"street_name": ""
				}
			],
			"legs": [],
			"details": {},
			"ascend": 0.0,
			"descend": 0.0,
			"snapped_waypoints": "smarHih}pBq_k@l@[?"
		}
	]
}

Map-matching --> Data-traffic-integration
In the graphhopper-traffic-data-integration website in the Data Format section at the bottom I can see we are using point and id, but how is that related to edge which is using by graphhopper?

With data-traffic-integration I think we have got some limitations. I’m afraid we cannot import new file in a runtime, when a requests are comming all the time to the graphhopper. Why we would like to do that is to change the speed data over a day (depends on rush hours).
One more doubt is a time needed for update traffic data, and if we can update the same graph information all the time, or we need to clean cache before every update.
I think the better option instead of using data-traffic-integration tool, to somehow prepare big table of edge/speed time depended table and and store all data.

I have no experience in those tools. For now it’s just research and to find any good way to go.
I would be gratefull for any responses or suggestions.

Powered by Discourse