Finished a project very analogous to this, with marine vessel traffic data (AIS mostly).

It's fun to visualize these patterns, but the business case soon ran out. People don't really care about absolute numbers. Diff is what matters the most.

Most of the analysis is "OK, how does this value compare with one week ago". Can't really do that in QGIS/ArcGIS... no matter how pretty the visualization.

In the end, time series always emerges victorious in capturing the right information/visualization ratio.

Delighted to see more experimentation with QGIS and real time data.

As an example of another approach that doesn’t need the intermediate csv file and has a much cleaner “refresh”, I made an adapter plugin that accesses real-time robot data via ROS: https://github.com/locusrobotics/qgis_ros

The ability to extend QGIS because of the C++ and Python APIs is immensely powerful.

Is QGIS the go-to app for visualization of position data like this?

Can it send commands back to the app from which it's receiving position data? Things like pause/play and adjusting the time being displayed?

In this case I interface with ROS in Python so yes it could just by wiring the Python together from both ends.

QGIS is kind of the go-to app if you want to do serious geospatial data analysis. As for _visualization_, I'm not sure. There's probably better options for many domains. For ROS there's RVIZ, but the ROS community's spatial analytical tools are fairly immature.

From what I gather it's pretty damn far from the typical use-case of a GIS tool but still a cool experiment. I have no idea how it could be useful like this either.

If you could however do some number crunching using historical data to display things like average air traffic density and routes, it might become useful.

That's great. Thanks for sharing. Would be very interesting to see robotics and geospatial application combination

No problem! Yes, I've been equal measures frustrated that I see so little geomatics utilized in ROS and excited to explore it.

It's a bit bewildering that there hasn't been any traction for the robotics community to use more GIS. It's a match made in heaven.

There's examples in the presentation linked in the git repo if you're interested.

I've started exploring GIS using QGIS and a PostGIS backend over the last few months. It's an absolutely fascinating topic and there is so much interesting geospatial data available to play with.

This article makes it clear that I've only scratched the surface of what you can do with it.

Besides institutional lock-in, why are people still paying for ArcGIS?

There is a lot of inertia behind it. Some companies will have thousands of ArcMap documents which need to be maintained.

And the ecosystem has value all on its own. No single product is the best. But taken together it is the least bad option. Particularly when you want the same team to be able to construct lots of fairly dis-similar pieces of technology.

Arcgis online works really well and is easily customized. I haven't been able to get qgis mobile to work with county shapefile data. Infact I can't even get the desktop version to create it's start icon script, so it's effectively unusable.

I have done a couple of large projects in agol and had some pretty serious issues.

  * Vector layers are fast but you cannot easily query attributes for a shape within the esri web apps.
  * Feature layers are dead slow for large datasets.
  * The API sometimes relies on you submitting form encoded JSON documents which may or may not actually do what you want on the backend.  
  * API error message are not always descriptive.
  * The API documentation is not great.  For example try finding the correct string formatting for a date.

It is a mass of different systems clustered together into a tightly coupled mess.

Because it's much better than qgis. It's like Photoshop vs gimp. Technically there is nothing you can do in one that you can't do in the other, but the same can be said of MS Paint...