Density of particles and particle displacement

[gerardllorach]

when looking at the map from far away, it is hard to distinguish the path of the particles (step of particles is to small). When zoom in a lot, the particles cover all screen.

Probably this is because the step of a particle is done in lat-long instead of pixels

[gerardllorach]

To work on the density->

• Calculate the number of data points on screen (use view box of map, then check range of long-lat and count number of points with data). The range should be extended by the distance between data points (in case that the data point is just in the edge).
• The space that a data point should be calculated as an area. Two strategies: area of a circle or area of a square. Makes more sense square area, as data points are adjacent and cover the grid.
• Stablish a relationship of density and number of particles

• Use the zoom of view to determine the step size of each particle in pixels:

  • Close zoom:

  • Reduce particle path vertices (less vertices, faster rendering). Use the step in lat-long and relate it to the distance between datapoints / whole distance of the path. It does not make sense to have 6 vertices over a path that moves very little and uses the same datapoint. Maybe check the zoom level? -> If very close one might see corners in the path.

  • Far away zoom:

  • Increase the step of the vertices in the path, as one cannot understand the currents from far away. Risk: the step skips datapoints and create paths that do not make sense. OR

  • Increase the number of path vertices. If we decrease the number of vertices for paths that move very little, increasing the number of vertices in other paths should not affect performance.

Try this for now and see.

[gerardllorach]

To work on the density->

• Calculate the number of data points on screen (use view box of map, then check range of long-lat and count number of points with data). The range should be extended by the distance between data points (in case that the data point is just in the edge).
• The space that a data point should be calculated as an area. Two strategies: area of a circle or area of a square. Makes more sense square area, as data points are adjacent and cover the grid.
• Stablish a relationship of density and number of particles

Done in e9ac2385b5a970152696c25b34aada2e8e14f119 .

Nevertheless, the density does not take into account the path (when zoom in, path takes all space, when far away, path almost takes no space). Maybe can be solved in next iteration.

• Use the zoom of view to determine the step size of each particle in pixels:

  • Close zoom:

  • Reduce particle path vertices (less vertices, faster rendering). Use the step in lat-long and relate it to the distance between datapoints / whole distance of the path. It does not make sense to have 6 vertices over a path that moves very little and uses the same datapoint. Maybe check the zoom level? -> If very close one might see corners in the path.

  • Far away zoom:

  • Increase the step of the vertices in the path, as one cannot understand the currents from far away. Risk: the step skips datapoints and create paths that do not make sense. OR

  • Increase the number of path vertices. If we decrease the number of vertices for paths that move very little, increasing the number of vertices in other paths should not affect performance.

Try this for now and see.