Introduction
IBM Connections 4.0 Metrics track events from Connections applications and provide reports that display information about how people use Connections. Beside the out of box reports that Metrics provide, user can create customized reports based on the events that Metrics captured and involve them into Metrics to fit the better business needs. This article will introduce the basic events list that Metrics track from Connections 4.0, and explain how to do the customized calculations based on these events during the report customization. The events that each Metrics standard report used will also be involved in this article.
Events list that Metrics track
The events that Metrics track can be found in the sheet 'Events List&Basic Reports' in the spread sheet file attached with this article. We have below columns for the events:
-
User Action : The action that user take in Connections UI
-
Description : The description for this action
-
Reports : The report name in Metrics that calculating this action
-
Source : The source application of this event
-
Item Type : The entry type that referred by this event
-
Event Type : The action type of the event, like 'create' or 'delete'
-
Unique User Event Type: The action type for calculating unique users that taking this action
-
Unique Content Event Type: The action type for calculating unique contents that affected by this action
Note: The 'Reports' column shows the basic report that calculated only from single basic event, for the reports that need abstract events calculation, we have put them in the sheet 'Calculated Reports' and 'Top Ranking Reports'.
Reports that need abstract events calculation
We can find the reports that need abstract events calculation in the sheet 'Calculated Reports' and 'Top Ranking Reports' in the spread sheet file attached with this article. We have below columns for the 'Calculated Reports':
-
Reports : The report name in Metrics
-
Source : The application that the report is calculating events for
-
Item Type : The entry type that referred by this event
-
Event Type : The action type of the event, like 'visit'
-
Unique User Event Type: The action type for calculating unique users
-
Description : The description for this report
-
Calculated Events : The events that were used in the report calculation
And provide following columns for the 'Top Ranking Reports':
-
Reports : The report name in Metrics
-
Scope : The category that the report belongs to
-
Description : The description for this report
-
Calculated Events : The events that were used in the report calculation
Create calculation from Metrics events
We can create customized calculation based on the events that Metrics captured, the following topics will discuss the way to do this. The audience for this section should have the basic knowledge of Cognos dimensional calculation.
Basic events calculation
For each event in the Metrics event list in the 'Events List&Baisc Reports' sheet, we can generally calculate three kinds of metrics:
-
Number of times that the event happen.
To calculate this kind of metrics, we need to provide below measure and dimensions:
MEASURE : EVENT_COUNT
SOURCE : use the value of 'Source' column
ITEM : use the value of 'Item Type' column
EVENT : use the value of 'Event Type' column
DATE : select category in DAY/WEEK/MONTH/YEAR level from 'DATE' dimension
Eg: We can use event 'activity.created' to calculate metrics 'Number of new activities create in year 2012'. The dimensions as below.
MEASURE : EVENT_COUNT
SOURCE : ACTIVITIES
ITEM : ACTIVITY
EVENT : CREATE
DATE : 2012
In Cognos, the calculation expression will like this : tuple( [ACTIVITIES] , [ACTIVITY] , [CREATE] , [EVENT_COUNT] , [2012] )
-
Number of unique users that trigger the event.
To calculate this kind of metrics, we need to provide below measure and dimensions:
MEASURE : DAILY_UNIQUE_USER / WEEKLY_UNIQUE_USER / MONTHLY_UNIQUE_USER
(The measure used here is based on the level in DATE dimension we select to calculate. For example, if we select DAY level, then use 'DAILY_UNIQUE_USER')
SOURCE : use the value of 'Source' column
ITEM : use the value of 'Item Type' column
UNIQUE EVENT: use the value of 'Unique User Event Type' column
DATE : select category in DAY/WEEK/MONTH level from 'DATE' dimension
Eg: We can use event 'activity.created' to calculate metrics 'Number of unique users create activities in month 10/12'. The dimensions as below.
MEASURE : MONTHLY_UNIQUE_USER
SOURCE : ACTIVITIES
ITEM : ACTIVITY
UNIQUE EVENT: USER_CREATE
DATE : 10/12
In Cognos, the calculation expression will like this : tuple( [ACTIVITIES] , [ACTIVITY] , [USER_CREATE] , [MONTHLY_UNIQUE_USER] , [10/12] )
-
Number of unique items that were affected by the action.
To calculate this kind of metrics, we need to provide below measure and dimensions:
MEASURE : DAILY_UNIQUE_CONTENT / WEEKLY_UNIQUE_CONTENT / MONTHLY_UNIQUE_CONTENT
(The measure used here is based on the level in DATE dimension we select to calculate. For example, if we select DAY level, then use 'DAILY_UNIQUE_CONTENT')
SOURCE : use the value of 'Source' column
ITEM : use the value of 'Item Type' column
UNIQUE EVENT: use the value of 'Unique Content Event Type' column
DATE : select category in DAY/WEEK/MONTH level from 'DATE' dimension
Eg: We can use event 'files.file.downloaded' to calculate metrics 'Number of unique files downloaded in month 10/12'. The dimensions as below.
MEASURE : MONTHLY_UNIQUE_CONTENT
SOURCE : FILES
ITEM : FILE
UNIQUE EVENT: CONTENT_DOWNLOAD
DATE : 10/12
In Cognos, the calculation expression will like this : tuple( [FILES] , [FILE] , [CONTENT_DOWNLOAD] , [MONTHLY_UNIQUE_CONTENT] , [10/12] )
Note: Not all events support these three kinds of calculations, check the event list table for details.
Blank value for column 'Unique User Event Type' or 'Unique Content Event Type' means this event does not support this kind of calculation.
If want to specify the calculation for a community, select the exact community in COMMUNITY dimension and add it to the calculation expression.
Abstract events calculation
Beside the basic events, we also have some abstract events that can be used to calculate some useful reports as below:
-
Number of unique followed people
MEASURE : DAILY_UNIQUE_USER / WEEKLY_UNIQUE_USER / MONTHLY_UNIQUE_USER
(The measure used here is based on the level in DATE dimension we select to calculate. For example, if we select DAY level, then use 'DAILY_UNIQUE_USER')
SOURCE : PROFILES
ITEM : PERSON
UNIQUE EVENT: USER_BE_FOLLOWED
DATE : select category in DAY/WEEK/MONTH level from 'DATE' dimension
-
Number of unique contributors
MEASURE : DAILY_UNIQUE_USER / WEEKLY_UNIQUE_USER / MONTHLY_UNIQUE_USER
(The measure used here is based on the level in DATE dimension we select to calculate. For example, if we select DAY level, then use 'DAILY_UNIQUE_USER')
SOURCE : Select the app from SOURCE dimension that you want to calculate.
ITEM : DEFAULT
UNIQUE EVENT: USER_CONTRIBUTE
DATE : select category in DAY/WEEK/MONTH level from 'DATE' dimension
COMMUNITY : DEFAULT for global Apps, select the exact community when calculating for a community.
-
Number of unique authenticated visitors
MEASURE : DAILY_UNIQUE_USER / WEEKLY_UNIQUE_USER / MONTHLY_UNIQUE_USER
(The measure used here is based on the level in DATE dimension we select to calculate. For example, if we select DAY level, then use 'DAILY_UNIQUE_USER')
SOURCE : Select the app from SOURCE dimension that you want to calculate.
ITEM : DEFAULT
UNIQUE EVENT: USER_LOGIN
DATE : select category in DAY/WEEK/MONTH level from 'DATE' dimension
COMMUNITY : DEFAULT for global Apps, select the exact community when calculating for a community.
MEASURE : EVENT_COUNT
SOURCE : Select the app from SOURCE dimension that you want to calculate.
ITEM : DEFAULT
EVENT : VISIT
DATE : select category in DAY/WEEK/MONTH level from 'DATE' dimension
COMMUNITY : DEFAULT for global Apps, select the exact community when calculating for a community.
Conclusion
After get familiar with the events calculation introduced in this topic, user can follow the report customization instructions on Connections 4.0 info center to begin the new report creation.
About the author
Xiao Feng Zhang is a Staff Software Engineer based at IBM's Beijing, China, Lab. He currently works on Cognos development for Connections Metrics. He can be contacted at zxfzhang@cn.ibm.com