Event Analysis
Event analysis is used to study the frequency of certain behavioral events. You can conduct multidimensional analysis of user behavior through custom metrics, groupings, filters, and various visual charts.
Use cases
Event analysis are commonly used when analyzing user behaviors, for example:
- Query on user usage of certain product functions (such as adding favorite, video playback, view live stream, etc.);
- Compare different group of users' behaviors, such as number of login per different country;
- Compare different channel's effectiveness, such as sign-up rate per traffic source.
Key concept:
- Metric: perform aggregation on a selected event, support the following metrics:
- User number: count number of event
- Event number: count number of distinct users
- County by: count distinct value of a specified dimension (string type)
- Max/Min/Sum/Avg/Median by: aggregate the values of a specified metric (number type)
How to use
- Select an event, and select the aggregation method for the metric.
- Add filter to the event by clicking the icon next to the metric.
- Select an event parameter or user attribute as filter. You can add multiple filters by clicking on the icon. You can also configure the filter relationship by clicking on
AND
orOR
. - Repeat above step to add more metric if needed.
- If needed, configure global filter by selecting event parameter or user attributes. Similar to event filter, you can add multiple global filters and configure the filter relationship.
- If needed, configure grouping by selecting an event parameter or user attribute
- Click on Query button to start the analysis.
- Adjust the data granularity, such as Daily, Weekly, Monthly, if needed.
- Adjust query time range if needed.
- Click on Save to Dashboard to save the analysis to a Dashboard, input a name, description, and select a dashboard and sheet.
Example:
Calculate the daily page views (PV - Page View) and active user count (UV - Unique Visitor) on the web from different countries over the past month, requiring active users to have a session duration of at least 30000 millisecond.
Steps
- Select the Event Analysis model.
- In the left Define Metrics area, choose
Page view
as the metric for calculating events and selectEvent number
as the metric type. - Click the + Add Event button to add another metric. Choose
App end
as the metric for calculating events and selectUser number
as the metric type. - Click the filter icon to the right of
App end
to add an event filter condition:- Filter property:
Session
/Session duration(msec)
- Operation:
>=
- Value:
30000
(the unit ofSession duration(msec)
is millisecond)
- Filter property:
- Configure a global filter in the right Filters area:
- Choose
Event preset
/Platform
as the filter property. - Operation:
=
- Value:
Web
- Choose
- In the right Attribute Grouping area, configure grouping by selecting
Geography
/Country
. - In the time selector at the bottom, choose
Last 1 Month
and click OK. - Click the Save to Dashboard button in the top right corner. In the pop-up dialog, enter:
- Chart Name:
PV and UV
- Chart Description:
PV and UV on the web over the past month (at least 30 seconds)
- Choose a Dashboard: Select a dashboard. (You need to create a dashboard first, follow the Create dashboard,)
- Choose a Worksheet: Select a worksheet.
- Click OK.
- Chart Name: