Data: Cloud Scenes for Huge Location Datasets

Cloud Scenes

State-of-the art hardware like e-obs 1Hz sensors easily produce millions of GPS fixes in the context of large-scale studies.

Cloud scenes are the Firetail’s approach to enable interactive exploration of huge datasets.

Firetail’s cloudscenes are indexed flavours of your data. Currently, we offer cloud scene processing and access as a service for large-scale studies on demand.

Contact us for details: Firetail Support

Opening a Cloud Scene

A local cloudscene can be opened by using File > Firetail Cloud > Open Firetail Cloud Scene. Associated reference-data will be included if available.

Another way to open a cloud scene is enter an ID via File > Firetail Cloud > Open Firetail Cloud Scene by ID. Given you entered a valid ID, the associated cloud scene will be downloaded and opened.

Note that cloud scenes comprise the complete dataset in an indexed fashion allowing for real-time analysis. The initial download, however, may still require several dozen Gigabytes of data.

Data Point View

After opening the cloud scene, all points are shown simultaneously. This is but an overview of the data. Yet, coloring by reference data or visual exploration is possible, like toggling individuals.

Selection

Use <shift>-drag to select a region of interest among the points. This will induce a window bracketing around the earliest time point among the selected data points. Use the R time bracket reset to remove the bracketing.

Firetails

The interval restriction dropdown menu provides you with a selection of time ranges (see Section Time Control). Moving individuals will then show a firetail following the selection time range of past movements. In this way it is possible to observe current movements in the context of recent patterns.

A video tutorial on how to work with Cloudscenes/Firetails: Tutorial

Short overview:

Toggle View

Use F1 to toggle the the point view. The selected individual’s positions are then shown for the current timestamp. Enable firetails of moderate length (1h, 3h) to get an impression of local paths that the individuals took.

Heatmaps

Use F2 to toggle a heatmap view of the point data. Firetail does count window-based point occurrences. These are visualized as a heatmap overlay with legend. The heatmap respects the current firetail selection and thus allows you to see where most of your observations are located.

Workflow

There is no ideal workflow considering the complexity of available data. Yet, we found that a fresh cloud scene could be explored comfortably using a combination of the following:

  1. Shift-drag to span a rectangular region across a set of data points to define a region of interest
  2. Press F1 to hide the point cloud
  3. Use Layers -> Sunrise to see night/day
  4. Select a firetail length of 1-2 hours to see trajectories
  5. Select a step length of 1h, use +/- to focus on time of day that contains interesting patterns
  6. Select a firetail length of 6-24 hours to see patterns spanning one day
  7. Select a step length of 1 day to step through days
  8. Found a region of interest?
    1. Decrease firetail & step length
    2. Choose a sensible replay speed and press Play

complex drilldown

Is my data available as a cloud scene?

We offer to convert your high-res (movebank) data to a cloud scene so you can make the most of your data. Contact us for details: support@firetail.de