[OSM-ja] experience with Navitime.co.jp coordinate system?

[田渕直]

Hi Tom,

Some googling concluded that:
   - The navitime's coords are ordinary lat/lng's, but
   - they are represented in *milli-second angles.*
(cf. http://ja.wikipedia.org/wiki/Wikipedia:%E5%9C%B0%E5%9B%B3%E3%82%B5%E3%82%A4%E3%83%88%E4%B8%80%E8%A6%A7#NAVITIME, etc.)

So, dividing them by 3600*1000:
	lat-or-lng / (3600*1000)
gives you familiar values in degrees.

For instance, the first line of your CVS:
     511368020,156819220,079-1123
transforms to:
     142.0466722222222,43.56089444444444
which points to a certain place in Hokkaido:
     https://goo.gl/maps/6n9zO
where, ignoring a slight gap bw. Tokyo Datum and WGS84, the postal code 079-1123 points:
     https://goo.gl/maps/bnXhT

hope it helps,

Noashi Tabuchi (tabuchi_naoshi ＠ trek.co.jp)

On 2015/02/25 4:06, Tom Lee wrote:
> I am still working to connect e-Stat data together into a shapefile of postal boundaries in Japan. I have found myself encountering data from services that use the navitime.co.jp <http://navitime.co.jp/> mapping platform. These coordinates seem to employ an unusual datum. There are some references on the web to it being the Tokyo Datum, and the constants present in some of the navitime code <https://gist.github.com/sbma44/49354581f45d3c7e10a9> indicate the Bessel Ellipsoid, which supports this idea.
>
> However, there are some strange aspects to these coordinates, such as multiplying them by constants like 1E3 * Math.PI / 648E3. I have been unable to successfully transform them into a known coordinate system. Does anyone have experience with the coordinate system used by the Navitime API?
>
> Here's a link to a CSV file of coordinates and post codes:
>
> http://cl.ly/1e413I1u1c12/navitime.csv
>
>
> _______________________________________________
> Talk-ja mailing list
> Talk-ja ＠ openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-ja
>