5 ms·

It would be nice if there was a way to manually enter location, since the location API on desktop is for me inaccurate by hundreds of miles.

CraftThatBlock 7y ago
You can spoof it in the inspector, under application/sensors I think
mkl 7y ago
Agreed. Actually, I think the default behaviour should be to show a map with a Voronoi diagram so you can immediately see the closest volcano for any location. Hm. The volcano locations are in a JSON list, and SciPy can do spherical Voronoi calculations... Edit: Someone already did it for Iceland (though without marking the volcano positions): https://www.flickr.com/photos/78506020@N00/6807458402/ https://www.flickr.com/photos/78506020@N00/6807458402/
- mkl 7y ago
  I worked out how to make matplotlib do it: https://imgur.com/lsxjRXP https://imgur.com/lsxjRXP
- rontoes 7y ago
  Do you have any experience with the Google Maps API? I would love to implement this.
  - mkl 7y ago
    No I don't. All you need, though, is the ability to draw straight lines (well, great circle lines) between points. You could do the region generation once offline, in Python: import numpy as np from scipy.spatial import SphericalVoronoi import json with open('volcanos.json', 'rt') as f: volcanos = json.load(f) #There are some nulls in the data: volcanos = [v for v in volcanos if v['longitude'] != None and v['latitude'] != None] def longlat_to_xyz(longlat): longlat = longlat.reshape((-1, 2))*(np.pi/180.) r = np.cos(longlat[:, 1]) return np.column_stack((r*np.cos(longlat[:, 0]), r*np.sin(longlat[:, 0]), np.sin(longlat[:, 1]))) #go through set to remove duplicates: volcano_longlat = np.array(list(set((v['longitude'], v['latitude']) for v in volcanos))) volcano_xyz = longlat_to_xyz(volcano_longlat) sv = SphericalVoronoi(volcano_xyz) sv.sort_vertices_of_regions() #make the vertices go around each shape #find unique edges (they're shared between regions): edges = set() for region in sv.regions: region.append(region[0]) for i in range(len(region)-1): edges.add((region[i], region[i+1])) long_min = -30. long_max = 330. def xyz_to_longlat(xyz, long_min=long_min, long_max=long_max): long = np.arctan2(xyz[:, 1], xyz[:, 0])*(180./np.pi) long[long<long_min] += 360. long[long>long_max] -= 360. return np.column_stack((long, np.arcsin(xyz[:, 2])*(180./np.pi))) vertices = xyz_to_longlat(sv.vertices) #To draw the regions on Google Maps etc., save edges and vertices as JSON and load into JS. #To draw the linked map with Matplotlib (very slow for hi-res, and it was fiddly setting up the right conda environment): import matplotlib.pyplot as plt from mpl_toolkits.basemap import Basemap fig, ax = plt.subplots(figsize=(16, 10)) m = Basemap(projection='robin', lon_0=-200, resolution='h', ax=ax) m.drawcoastlines(linewidth=.1) m.drawcountries(linewidth=.1) volcano_longlat2 = xyz_to_longlat(volcano_xyz) x, y = m(volcano_longlat2[:, 0], volcano_longlat2[:, 1]) m.plot(x, y, 'ro', markersize=.5) for edge in edges: m.drawgreatcircle(vertices[edge[0], 0], vertices[edge[0], 1], vertices[edge[1], 0], vertices[edge[1], 1], color='b', linewidth=.3) fig.tight_layout() fig.savefig('volcano_voronoi.png', dpi=600)
    - bonoboTP 7y ago
      You have to use "is not None" instead of "!= None"
manojlds 7y ago
Looks off even on mobile with location enabled.
ComputerGuru 7y ago
Where I am, it seems it could be off hundreds of miles in any direction and I’d still be far away from any volcano whatsoever.
sleavey 7y ago
Or for those who disable the location requests. I never give websites permission to see my location (not that it's very useful on desktops without GPS). Almost all websites fall back to a search box when they can't get the location automatically.
mobilemidget 7y ago
For me it doesn't seem to return anything via the API, just says calculating indefinitely
- rontoes 7y ago
  Which browser/device are you using?
  - egfx 7y ago
    iOS SE safari
  - mobilemidget 7y ago
    catalina safari
rontoes 7y ago
I will add this feature!

CraftThatBlock 7y ago

You can spoof it in the inspector, under application/sensors I think

mkl 7y ago

Agreed. Actually, I think the default behaviour should be to show a map with a Voronoi diagram so you can immediately see the closest volcano for any location. Hm. The volcano locations are in a JSON list, and SciPy can do spherical Voronoi calculations... Edit: Someone already did it for Iceland (though without marking the volcano positions): https://www.flickr.com/photos/78506020@N00/6807458402/ https://www.flickr.com/photos/78506020@N00/6807458402/

mkl 7y ago
I worked out how to make matplotlib do it: https://imgur.com/lsxjRXP https://imgur.com/lsxjRXP
rontoes 7y ago
Do you have any experience with the Google Maps API? I would love to implement this.
- mkl 7y ago
  No I don't. All you need, though, is the ability to draw straight lines (well, great circle lines) between points. You could do the region generation once offline, in Python: import numpy as np from scipy.spatial import SphericalVoronoi import json with open('volcanos.json', 'rt') as f: volcanos = json.load(f) #There are some nulls in the data: volcanos = [v for v in volcanos if v['longitude'] != None and v['latitude'] != None] def longlat_to_xyz(longlat): longlat = longlat.reshape((-1, 2))*(np.pi/180.) r = np.cos(longlat[:, 1]) return np.column_stack((r*np.cos(longlat[:, 0]), r*np.sin(longlat[:, 0]), np.sin(longlat[:, 1]))) #go through set to remove duplicates: volcano_longlat = np.array(list(set((v['longitude'], v['latitude']) for v in volcanos))) volcano_xyz = longlat_to_xyz(volcano_longlat) sv = SphericalVoronoi(volcano_xyz) sv.sort_vertices_of_regions() #make the vertices go around each shape #find unique edges (they're shared between regions): edges = set() for region in sv.regions: region.append(region[0]) for i in range(len(region)-1): edges.add((region[i], region[i+1])) long_min = -30. long_max = 330. def xyz_to_longlat(xyz, long_min=long_min, long_max=long_max): long = np.arctan2(xyz[:, 1], xyz[:, 0])*(180./np.pi) long[long<long_min] += 360. long[long>long_max] -= 360. return np.column_stack((long, np.arcsin(xyz[:, 2])*(180./np.pi))) vertices = xyz_to_longlat(sv.vertices) #To draw the regions on Google Maps etc., save edges and vertices as JSON and load into JS. #To draw the linked map with Matplotlib (very slow for hi-res, and it was fiddly setting up the right conda environment): import matplotlib.pyplot as plt from mpl_toolkits.basemap import Basemap fig, ax = plt.subplots(figsize=(16, 10)) m = Basemap(projection='robin', lon_0=-200, resolution='h', ax=ax) m.drawcoastlines(linewidth=.1) m.drawcountries(linewidth=.1) volcano_longlat2 = xyz_to_longlat(volcano_xyz) x, y = m(volcano_longlat2[:, 0], volcano_longlat2[:, 1]) m.plot(x, y, 'ro', markersize=.5) for edge in edges: m.drawgreatcircle(vertices[edge[0], 0], vertices[edge[0], 1], vertices[edge[1], 0], vertices[edge[1], 1], color='b', linewidth=.3) fig.tight_layout() fig.savefig('volcano_voronoi.png', dpi=600)
  - bonoboTP 7y ago
    You have to use "is not None" instead of "!= None"

manojlds 7y ago

Looks off even on mobile with location enabled.

ComputerGuru 7y ago

Where I am, it seems it could be off hundreds of miles in any direction and I’d still be far away from any volcano whatsoever.

sleavey 7y ago

Or for those who disable the location requests. I never give websites permission to see my location (not that it's very useful on desktops without GPS). Almost all websites fall back to a search box when they can't get the location automatically.

mobilemidget 7y ago

For me it doesn't seem to return anything via the API, just says calculating indefinitely

rontoes 7y ago
Which browser/device are you using?
- egfx 7y ago
  iOS SE safari
- mobilemidget 7y ago
  catalina safari

rontoes 7y ago

I will add this feature!