This is a migrated thread and some comments may be shown as answers.

RadMap VisualizationLayer Cluster Performance

4 Answers 168 Views
Map
This is a migrated thread and some comments may be shown as answers.
Thomas
Top achievements
Rank 1
Iron
Iron
Iron
Thomas asked on 05 Feb 2018, 05:33 PM

Hello,

We have an issue about the VisualizationLayer control fo RadMap, we are currently on version 2016 R3

We try to display +1500 data in a country with smooth performance when moving or zooming, but actually we tried a lot of things and nothing change.

 

We started to use InformationLayer and it was perfectly smooth but all points were on top of each other and it was ugly.

So we change to VisualizationLayer, while setting ItemsSource + Clustering enabled:

- Dataloading was correct

- Moving was smooth (in same zoom)

- But Zooming in or out was extremely slow and froze everything for some seconds

 

Then we added GenerateClustersOnZoom = false

It helps a little when its zooming but at the end we have the same freeze as before when it actually refresh all clusters

And let some clusters on the map (I don't know why)

 

We also tried with a VirtualizationSource, it was a little better but still far from the behavior of the InformationLayer

 

What do we need to do to have Clustering as fast as the InformationLayer in term of moving/zooming on the map ?

Or why the clustering is so time consuming ?

 

Regards

4 Answers, 1 is accepted

Sort by
0
Petar Mladenov
Telerik team
answered on 08 Feb 2018, 10:02 AM
Hello Thomas,

When extensive zooming is performed a lot of (Items.Remove) and (Items.Add) operations are scheduled and later performed. We are aware of such performance issues in VisualizationLayer with clustering enabled. That is why we highly encourage you to use this feature in combination with virtualization. For best results, these two can be also combined with AsyncReader which will add the elements in map asynchronously. You can check the Virtualization And Grouping example (it is available in our WPF demos too). Notice the items disappear before zooming (before refresh) which additionally speeds up performance. In your case probably the biggest effort will be organizing your data in SHP file. You can for example use QGIS. I hope this is an option that you will consider.

Regards,
Petar Mladenov
Progress Telerik
Want to extend the target reach of your WPF applications, leveraging iOS, Android, and UWP? Try UI for Xamarin, a suite of polished and feature-rich components for the Xamarin framework, which allow you to write beautiful native mobile apps using a single shared C# codebase.
0
Thomas
Top achievements
Rank 1
Iron
Iron
Iron
answered on 08 Feb 2018, 10:26 AM

Hello,

Thanks you for the reply.

We will try creating a SHP file while using AsyncReader as the example shows.

I will tell you if it works fine.

 

Regards

0
Thomas
Top achievements
Rank 1
Iron
Iron
Iron
answered on 08 Feb 2018, 01:21 PM

I was thinking about my problem and your answer:

Using a shape file will not resolve my problem of zooming ?

I already use the virtualization of the source to tile the points and clusters display but zooming is very slow and freeze everything until all tiles are loaded. That's the main part I want to optimize

Do you have a solution for this problem ?

 

Thank you

0
Petar Mladenov
Telerik team
answered on 09 Feb 2018, 07:47 AM
Hello Thomas,

I mentioned the root cause of the zooming performance - lots of objects removed and re-added to a Canvas. I then suggested the combination of shape file, async reader  and virtualization as the best known optimization for such big amount of objects and clusters. I hope it will give you the best results in your scenario. Another thing worth trying is stopping the animations - set UseSpringAnimations = False on RadMap.

Regards,
Petar Mladenov
Progress Telerik
Want to extend the target reach of your WPF applications, leveraging iOS, Android, and UWP? Try UI for Xamarin, a suite of polished and feature-rich components for the Xamarin framework, which allow you to write beautiful native mobile apps using a single shared C# codebase.
Tags
Map
Asked by
Thomas
Top achievements
Rank 1
Iron
Iron
Iron
Answers by
Petar Mladenov
Telerik team
Thomas
Top achievements
Rank 1
Iron
Iron
Iron
Share this question
or