Module talk:Location map/data/China Tibet

Tests

Other:

39.91, bottom = 25.82, left = 77.61, right = 103.47

Test 1

class=notpageimage|

Nepal/India/China border at 30.197206°N 81.034478°E / 30.197206; 81.034478

Test 2

class=notpageimage|

Bhutan/India/China border at 27.327338°N 88.921275°E / 27.327338; 88.921275

Test 3

class=notpageimage|

Xinjiang/Qinghai/Tibet border at 36.08684°N 89.74457°E / 36.08684; 89.74457

Test 4

class=notpageimage|

Sichuan/Qinghai/Tibet border at 32.53755°N 97.70142°E / 32.53755; 97.70142

map edge coordinates

Hi, User:Himalayan_Explorer asked me to see if I could help with this template, as I'd done the original map. I've been looking at the original sources I used, and these are the numbers I've come up with:

top = 41.1

bottom = 26.4

left = 72

right = 104

That's based on what the lat/long is at the centre of the map, but I'm not sure if that's the right way to do it. The centre of the map should work out to be the two connected lakes on the border between Qinghai and the TAR. It looks like that numbers being used now are working so I'll just leave my calculations as a note here. I'm glad to see this map is getting used! --Keithonearth (talk) 18:51, 13 October 2009 (UTC)

I tried your suggested coordinates, and all four test cases were further away than what they are right now. If you have a set of coordinates and pixel locations for at least two points on the map (the more the better), I can perform a least squares fit to get the best possible choice for the limits of the map. I think the idea is to try to make points within the region of Tibet as close as possible, but not worry so much about points outside of Tibet. Plastikspork _―Œ^(talk) 19:02, 13 October 2009 (UTC)

NNW raised concerns about the actual projection of the map. He's the expert on maps on wiki, it would seme to be the projection rather than the coordinates causing the problem... Himalayan 19:04, 13 October 2009 (UTC)

Himalayan Explorer's coordinates (top = 41.2 bottom = 25.5 left = 77.5 right = 103.3) seems to be a reasonable compromise, checking the test cases above. Plastikspork _―Œ^(talk) 20:02, 13 October 2009 (UTC)

Here's some pixel and lat/long coordinates for a few points on the map. If you would like I can give more. Lat/long coordinates are only accurate give or take 0.1 deg. pix are rounded to the nearest pix.

• TAR/Xinjiang/Ladakh border junction:

x:126, y:627

34.4N, 79E

• India/TAR/Nepal(west):

x:201, y:329

30N, 81E

• Yunnan/Sichuan/TAR:

x:1269, y:265

29N, 99E

• Xinjiang/Qinghai/Tar:

x:723, y:708

36.2N, 89.8E

--Keithonearth (talk) 20:19, 13 October 2009 (UTC)

Based on those four points, the best (least-squares) fit is (top = 41.5 bottom = 24.8 left = 77.2 right = 103.4), which is pretty close to what Himalayan Explorer has in there right now. Thanks! Plastikspork _―Œ^(talk) 20:32, 13 October 2009 (UTC)

I've just uploaded a new version of File:Tibet&neighbors.svg that includes latitude and longitude lines (identical in all respects to File:Tibet&neighbors Locator map.svg other than the labels, and lines). The lines are faint, and don't seem to show at all at less than 900px. My idea was they will provide more info about the projection, and see if that is what is causing the push-pins locators to render inaccurately. It very well be my error in making the map, but I hope there's a work around. Even if the lines don't help in that respect I think they may a useful addition to the map. --Keithonearth (talk) 22:01, 14 October 2009 (UTC)

Your map has got an azimuthal or conic projection (didn't check which one) but location maps need an equirectangular projection to work properly. Mercator looks similar but it's wrong. Other projections are possible but then the code has to be completely different, see here the French template for Canada. Maps with an equirectangular projection look quite strange so in the end they get stretched in north/south direction:

• 
  UK in normal equirectangular projection
• 
  UK after 170 % N/S stretching

For Tibet a 115 % stretching would be best. This results from 100 * (1/cos longitude of the centre of the map). UK got 170 % stretched with 54° N (that's York) supposed as the central longitude. NNW (talk) 08:43, 15 October 2009 (UTC)

Thanks for the info NNW, it really clarifies what's going wrong. I just was looking at the map in inkscape, and it's past my abilities to make that sort of modification to the map, if inkscape is indeed able to do it at all. If someone knows more than me, go for it, and if there's any questions I can help with I'll do my best to answer. --Keithonearth (talk) 18:20, 15 October 2009 (UTC)

Merging of this template with that for Tibet?

Both templates (Tibet, China Tibet) point to the same map and region, so shouldn't they be merged? One being redirected to the other? C1MM (talk) 05:38, 22 November 2020 (UTC)