I’m encountering some challenges with the Points on Map visual in QuickSight and could use some help understanding or resolving them. Here’s the situation:
Dataset and Testing
My original dataset contains many records, with fields such as session_id, longitude, latitude, os, browser, category, etc. However, in the Points on Map visual:
Why does my original dataset show 25k points, but removing the session_id from the color attribute reduces this to 10k points? Is QuickSight applying some kind of grouping, sampling, or optimization that explains this behavior?
Is there a way to represent the data as a percentage of the total points? Is there a way to configure the map to show percentages of the total instead of raw counts for clusters? Alternatively, is there a better way to represent this type of data given the limitations of the Points on Map visual?
Why do different IDs with the same location only show 5 points? In session_data_with_same_location_30k, I expected some aggregation (e.g., count per location) to be displayed, but it seems only a small subset of the data appears. Is there a way to explicitly configure the map to handle this?
Handling the 10k Limit with Reduced Precision? Are both Basic points and Cluster points subjected to the 10k limit? Could I reduce the precision of longitude and latitude (e.g., truncate or round coordinates) and then show just the session’s count and use it as the dot size?
Reduction in Points when Removing session_id:
The reduction in the number of points when removing the session_id from the color attribute is likely due to QuickSight's optimization and aggregation behavior. QuickSight may be applying some form of clustering or aggregation to the data when the number of unique data points exceeds a certain threshold. By including the session_id in the color attribute, QuickSight is able to display more individual data points, as it can differentiate them based on the unique session identifier. When you remove the session_id, QuickSight may be consolidating or aggregating the data points, resulting in a lower total number of displayed points.
Representing Data as Percentage of Total:
To represent the data as a percentage of the total points, you can create a calculated field that calculates the percentage for each data point. For example:
ROUND(COUNT({session_id}) / SUM(COUNT({session_id})) OVER() * 100, 2)
This will give you the percentage of the total points for each data point. You can then use this calculated field as the size or color attribute in the Points on Map visual.
Alternatively, you can configure the map to show percentages of the total instead of raw counts for clusters. To do this, go to the visual's properties, find the "Cluster Options" section, and enable the "Show Percentage" option.
Handling Data with Same Location:
The behavior you're seeing, where only a small subset of the data with the same location is displayed, is likely due to QuickSight's handling of overlapping data points. When there are multiple data points with the same location, QuickSight may choose to display only a limited number of them to avoid visual clutter.
To address this, you can try the following:
Experiment with the "Cluster Options" settings, such as adjusting the "Cluster Radius" or enabling "Cluster by Location" to see if it improves the display of the data points.
Consider using a different visual type, such as a Choropleth map, which may be better suited for displaying aggregated data at the location level.
Handling the 10k Limit with Reduced Precision:
Both Basic points and Cluster points are subject to the 10k limit in QuickSight. To work around this, you can try reducing the precision of the longitude and latitude coordinates by truncating or rounding them. This can help reduce the number of unique data points and potentially allow you to display more data within the 10k limit.
Here's an example of how you can create a calculated field to reduce the precision of the coordinates:
ROUND({latitude}, 2) AS latitude_rounded,
ROUND({longitude}, 2) AS longitude_rounded
You can then use these rounded coordinates in your Points on Map visual and adjust the dot size based on the count or percentage of the data points.
Keep in mind that reducing the precision of the coordinates may result in some loss of detail, so you’ll need to find the right balance between visual clarity and data representation.
The first thing I was not sure of, is why with my original dataset Quicksight can display 25k points although the limit is 10k, Do you have any thoughts?
Every session_id is unique, how can I aggregate the cluster location points to be the percentage of the total?
I was not able to find anything like this, can you please share where to find those?
Unfortunately, this needs to be done on the data level since it cannot be done in a calculated field, “Geospatial field longitude isn’t supported in a calculated field.”
If there is a better way to handle this, please also let me know.
Since we have not heard back from you, I’ll go ahead and close this topic. However, if you have any additional questions, feel free to create a new topic in the community and link this discussion for relevant information.
