Datapoints limit of syncfusion charts

[Dhananjay Ghongane]

Do we have any documentation available for limit of number of datapoints in syncfusion charts ?

For Example on macbook pro chrome browser :

- Fast line chart can plot  10 thousand datapoint having numeric/datetime x-axis across series
- Fast line chart can plot  5 thousand datapoint having categorical x-axis across series
- Circular charts can handle 1k data points

- Bar charts will handle 1k data points 
- Scatter chart will handle 10 k datapoints 

We need it to decide machine configuration if that affects.
limits will change on desktop vs mobile app vs mobile browser ?

NOTE 1: This is not regarding how to improve performance, We went through other forum questions and we applied them.
NOTE 2:  These performance stats are important for us to avoid browser hang / mobile application hang issues at customer site because of more number of datapoints.  If we know performance stats we will trim data OR show warnings before hand OR do some aggregation before plotting if number of datapoints are out of capacity.  

[Lokesh Palani]

Hi Dhananjay,

We apologize for the delay in response, but we have investigated your query thoroughly and have some recommendations for you.

Based on the device configuration and RAM, the exact data point count for optimal performance may vary. However, we suggest the following features and improvements to avoid crashes:

1. We do not recommend using the CategoryAxis for more data points. Running SfCartesianChart with 1,000 data points in the CategoryAxis leads to performance issues because it is index-based rather than range-based, causing every data point to be iterated.

2. For circular charts, we have grouped features to avoid small data values or more data points for better visualization. Please refer to the below Ug link for more information.

https://help.syncfusion.com/flutter/circular-charts/chart-types/pie-chart#grouping-data-points

3. We suggest avoiding loading more data points during the loading time. We recommend the use of lazy loading to avoid crash issues. Please refer to the below KB links for more information:

https://support.syncfusion.com/kb/article/10855/how-to-lazily-load-more-data-to-the-chart-sfcartesianchart

https://support.syncfusion.com/kb/article/11533/how-to-synchronize-on-demand-loading-in-the-cartesian-charts-sfcartesianchart

I hope these recommendations will be helpful in avoiding crash issues. Please let us know if you have any further queries or concerns.

Thank you for choosing Syncfusion.

Regards,

Lokesh.

[Dhananjay Ghongane]

Hello Lokesh,

Thank you for your response.

Defiantly performance improvement suggestion will help (We saw it in other forum answers as well).
Lazy loading is not the option in our case because we want entire data view in one glance to look after data patterns.

"Based on the device configuration and RAM, the exact data point count for optimal performance may vary." 
=> Could you provide stat for any particular device with disclaimer that it may vary. 

E.g 
Device 1: 2.6 GHz Dual-Core Intel Core i5.  and  8 GB 1600 MHz DDR3  Mac Book pro:

Optimal Stats : 

• Fast line chart   --------- datapoint having numeric/datetime x-axis across series
• Fast line chart ---------  datapoint having categorical x-axis across series
• Circular charts --------- data points
• Bar charts ---------  data points 
• Scatter chart ---------  datapoints

Device 2:  some other mobile device and its configuration.

Optimal Stats : 

• Fast line chart   --------- datapoint having numeric/datetime x-axis across series
• Fast line chart ---------  datapoint having categorical x-axis across series
• Circular charts --------- data points
• Bar charts ---------  data points 
• Scatter chart ---------  datapoints

Could you please fill in the blanks. With these referral example i will try to set high level (not exact) user expectations of our app

Regards,
Dhananjay.

[Lokesh Palani]

Hi Dhananjay,

This is the optimal performance of the device when using our chart.

For 1k data points

• Fast line chart   - datapoint having numeric (10 ms) / datetime x-axis across series(15ms)
• Fast line chart -    datapoint having categorical x-axis across series (23 ms)
• Circular charts - data points (65 ms)
• Bar charts - data points (59 ms)
• Scatter chart - datapoints (31  ms)

For 5k data points

• Fast line chart   - datapoint having numeric (17 ms) / datetime x-axis across series (24 ms)
• Fast line chart - datapoint having categorical x-axis across series (31 ms)
• Circular charts - data points (97 ms)
• Bar charts - datapoints (73 ms)
• Scatter chart - datapoints (43 ms)

We are still improving the performance.

Regards,

Lokesh.

[Dhananjay Ghongane]

Lokesh,

Thanks for sharing stats.

Could you also share on which device you took these reading ?

Also please share same stats for 7k and 10k datapoints.


Reagrds,
Dhananjay. 

[Lokesh Palani]

Hi,

Sorry for the miscommunication. We had shared incorrect information, but we are currently working on improving performance in the chart. We have now provided the correct details based on your requirement. We plan to include this improved performance in our upcoming releases, which are expected in October 2023. We appreciate your patience until then. Please let us know if you need any further assistance.

The below performance details were recorded on an Android mobile device with 12 GB of RAM and a Mediatek Dimensity 1200 processor.

For 1k data points

• Fast line chart   --------- datapoint having numeric (151 ms) / datetime x-axis across series (163 ms)
• Fast line chart --------- datapoint having categorical x-axis across series (210 ms)
• Circular charts --------- data points (195 ms)
• Bar charts --------- data points (173 ms)
• Scatter chart --------- datapoints (164 ms)

For 5k data points

• Fast line chart   --------- datapoint having numeric (165 ms) / datetime x-axis across series (171 ms)
• Fast line chart --------- datapoint having categorical x-axis across series (236 ms)
• Circular charts --------- data points (​277 ms)
• Bar charts --------- data points (213 ms)
• Scatter chart --------- datapoints (244 ms)

For 7k data points

• Fast line chart   --------- datapoint having numeric (181 ms) / datetime x-axis across series (193  ms)
• Fast line chart --------- datapoint having categorical x-axis across series (265 ms)
• Circular charts --------- data points (323 ms)
• Bar charts --------- data points (241 ms)
• Scatter chart --------- datapoints (301 ms)

For 10k data points

• Fast line chart   --------- datapoint having numeric (193 ms) / datetime x-axis across series (205 ms)
• Fast line chart --------- datapoint having categorical x-axis across series (281 ms)
• Circular charts --------- data points (384 ms)
• Bar charts --------- data points (256 ms)
• Scatter chart --------- datapoints (342 ms)

Macbook pro browser:

For 1k data points

• Fast line chart   --------- datapoint having numeric (1018 ms) / datetime x-axis across series (1020 ms)
• Fast line chart --------- datapoint having categorical x-axis across series (1019 ms)
• Circular charts --------- data points (2500 ms)
• Bar charts --------- data points (1019 ms)
• Scatter chart --------- datapoints (1020 ms)

For 5k data points

• Fast line chart   --------- datapoint having numeric (1021 ms) / datetime x-axis across series(1019ms)
• Fast line chart --------- datapoint having categorical x-axis across series (1020 ms)
• Circular charts --------- data points (2670 ms)
• Bar charts --------- data points (1022 ms)
• Scatter chart --------- datapoints (1022 ms)

Regards,

Lokesh.

[Dhananjay Ghongane]

Hello Lokesh,

Thank you for information. 

Just for my better understanding i have one question. 
Lets take below example from your stat.

Macbook pro browser: 

For 5k data points

• Bar charts --------- data points (1022 ms)

Does this mean:  Bar chart will load  5 thousand datapoints in 1022 milliseconds on macbook pro browser ?

[Lokesh Palani]

Hi ,

We have obtained the performance details five times and calculated the mean value. According to the shared details, Bar charts will take 1022 milliseconds to load 5000 data points. Please inform us if you require any further information.

Regards,

Lokesh.

[Dhananjay Ghongane]

Lokesh,

Thank you for all your findings. This will really help us. 

Please inform us once performance improved release is rolled out in October

Regards,
Dhananjay.

[Lokesh Palani]

Hi Dhananjay,

Most Welcome. Kindly get back to us if you have further queries. We are always happy to assist you.

We will let you know once the release has been rolled out, and we appreciate your patience until then.

Regards,

Lokesh.

[Lokesh Palani]

Hi Dhananjay,

Sorry for the inconvenience, we missed to notify the release details here. We are glad to announce that our Essential Studio 2023 Volume 4 release v24.1.41 was rolled out and is available for download under the following link.

Essential Studio 2023 Volume 4 Main Release v24.1.41 is available for download | Announcements Forums | Syncfusion

In the latest chart package version, we did some performance improvement, so we kindly request you to upgrade to the below mentioned version.

Version link,

https://pub.dev/packages/syncfusion_flutter_charts/versions/24.1.41

Please try the above-mentioned chart version and let us know the result.

We thank you for your support and appreciate your patience in waiting for this release. Please get in touch with us if you would require any further assistance.

Regards,         

Lokesh P.

[Dhananjay Ghongane]

Thank you for your reply, We will definitely check at our end.  

[Preethika Selvam]

Hi Dhananjay,

Most Welcome. Kindly get back to us if you have further queries. We are always happy to assist you.

Regards,

Preethika Selvam.