Data SaaS Platform usage documentation（3.0）

一、Platform address

Address：https://demo.datahanding.com/

二、Dashboard

1. The meaning of Dashboard

Dashboard is a collection of charts, charts include reports, the platform can support a report into multiple data Dashboard, can also support multiple reports into a data Dashboard.

2.A page overview of the Dashboard

The data Dashboard consists of three parts: the Dashboard directory, Dashboard configuration area, and data display area

• Dashboard directory: Create Dashboard, view self-built kanban or team members’ shared Dashboard
• Dashboard configuration area: includes adding Dashboard content, setting Dashboard sharing, adjusting Dashboard Settings, refreshing Dashboard, and setting global time filtering for Dashboard reports
• Data display area: Displays the information of each Dashboard chart, customize the size and chart display type

3. Data Dashboard usage guide

3.1 Create/Edit Dashboard

3.1.1 Create

Company executives, administrators, analysts, or other members who have the permission to create Dashboard can click [+] in the left column of Dashboard and select “New Dashboard”.

3.1.2 Rename

Mouse [right click] self-built Dashboard, select “rename”, you can rename the Dashboard.
Note: Shared with Me is a preset section of the system and does not support renaming operations.

3.1.3 Delete

Mouse [right click] self-built Dashboard, select “Delete Dashboard”, you can delete Dashboard.

3.1.4 Dashboard movement

Mouse [right click] self-built Dashboard, select “Move to”, you can move to the space or my Dashboard, you can also click “management”, Dashboard batch processing.

3.2 Manage Dashboard content

Click the button at the top right of the Dashboard and click the Settings button.

3.2.1 Add chart

Click [+] on the right to add the report

If there is no report, you can create one in Report Management or Behavior Analysis

· Report management

· Behavior analysis

3.3 Dashboard sharing

The platform supports sharing self-built Dashboard with other members of the team to jointly view Dashboard and achieve collaboration among team members.

3.3.1 Set Dashboard sharing

Click the “Share” button on the upper right of the Dashboard to enter the Dashboard sharing Settings

By default, Dashboard is visible only to yourself. Click [+] to add the shared object of Dashboard.

3.4 Duplicate Dashboard

In the process of using Dashboard, if you want to make further analysis and adjustment on the basis of a Dashboard, you can use the copy Dashboard function. By default, the supermanager, administrator, and analyst have the operation rights to copy Dashboard.
Mouse [right click] Dashboard, select “Copy Dashboard”, you can copy Dashboard.
· Copying the Dashboard will copy all the charts in the Dashboard at the same time
· The shared status of the copy Dashboard is “visible only to yourself” by default.

3.5 Global filtering

3.5.1 Modified analysis time

You can change the time frame in any state. The data granularity is day.
After selecting the time, execute the command in the following order：
1、To change the display content of the global time frame
2、After the calculation is complete, the time is changed to the same as the global time box and the report results are displayed

3.5.2 Scenario in which the time frame is modified

1、When the overall time frame is adjusted, the time frame of each Dashboard report on the page is adjusted temporarily and the result is re-calculated
2、Modifying the content of a Dashboard report and recalculating the report does not affect the overall filtering conditions
3、When you log out or switch Dashboard, the time box is cleared
4、When refreshing Dashboard, the time box options are not cleared

3.6 Bulk download

Download the corresponding chart data in Dashboard in batches. Each chart corresponds to a csv file.

4. Guidelines for the use of project space

4.1 Project space usage scenarios

After you set up Dashboard sharing under [My Dashboard], the Dashboard will be displayed under [My Dashboard] → [Shared to me]. With the deepening of the use of the system, the number of Dashboard in the project is increasing. At the same time, it brings difficulties to the Dashboard creator to set up Dashboard sharing, the sharing status management of Dashboard, the classification and sorting of the shared Dashboard by the shared people, and the Dashboard collaboration among team members. Using the function of project space, we can realize the batch sharing of Dashboard and the unified classification and sorting of Dashboard shared by many creators. Achieve the unity and structure of team Dashboard view, so that the communication and cooperation between team members is more efficient and convenient.

4.2 Create/Edit space

4.2.1 Create space

You can click [+] in the left column of Dashboard and select “New Space”. By default, the administrator can create space.

Select a space member and set the space member permissions in the Create Space TAB. In-space Dashboard is visible to space members by default.

4.2.2 Edit space

Mouse click […], select “Edit space”, you can edit the space.

The space creator can edit the space Settings, including modifying the space member list and changing the space name

4.2.3 Delete space

Mouse click […], select “Delete space”, you can delete the space.
Deleting a space does not delete the Dashboard directly, after deleting, the Dashboard under the space will be moved to the Dashboard creator [My Dashboard] folder.

4.3 New space signage

4.3.1 Create a new Dashboard directly into the space

Click [+] in the upper right corner of the left column of the Dashboard and select “New Dashboard” to save the Dashboard directly to the space where it is located.

4.3.2 Move self-built Dashboard to space

Click the Dashboard […], select “Move to”, you can directly move the Dashboard to the space.
Note: Moving the Dashboard does not change the sharing Settings of the Dashboard (i.e. the sharing status in the upper right corner [Dashboard sharing Settings] does not change with the movement of the Dashboard)

5. Dashboard management

5.1 Batch move or delete your own Dashboard

Click on the upper right corner of the left column of Dashboard [Dashboard management] to open the Dashboard management pop-up window, which supports batch movement or deletion of self-built Dashboard.
· Only batch management of self-built Dashboard is supported
· You can batch move your self-built Dashboard to “My Dashboard” or a project space with permissions

6. Dashboard chart

6.1 Dashboard chart composition

The Dashboard core area consists of multiple Dashboard charts (reports), and the form of the charts in the Dashboard is not completely consistent with that in the model.
Any Dashboard chart consists of a palette and chart content, and depending on the chart type (source), the options supported by the palette vary.

6.2 Operation rights of different roles on the Dashboard chart

三、Analysis

Event It represents a certain or a series of meaningful behaviors of users, such as registering an account, purchasing gift packages, etc. Through behavioral analysis, the real use process of users can be restored, which is an important part of game data analysis. For different analysis scenarios, this platform provides a variety of analysis models, you can choose according to the actual situation.

Events is the most basic model, which can calculate the aggregated metrics of user’s specific behaviors over a period of time, or the trend in metrics, such as whether the number of daily active uses (DAU) remains steady, how much today’s revenue is, etc.

Retention analyzes the retention rate, one of the KPIs. By selecting the trigger and return actions, you can get the new user’s retention rate at the second, third or seventh day easily. You can also calculate the LTV and ROI by “Calculate another metric”.

Funnel is used to analyze the number and percentage of users who complete specified steps in order. This enables you to quickly know the conversion or churn rate at specific steps, such as points to leave newbie guide or level stay data, and find possible problems.

Distribution divides intervals based on user participation. You can divide by the frequency of participation, number of days, etc., or by the sum of specific property of each user, such as the cumulative payment amount, and view the number and percentage of users in each interval.

Flows is generally used for exploratory analysis. You can visualize the inflow and outflow of users before and after key points through Sankey diagrams, and analyze user’s behavioral preferences, such as which activities users engage in first after logging in each day, or the key play pattern used by users before their loss.

Interval can analyze the conversion duration between two events in causal relationship, such as the median conversion time spent from registration to first payment, or the distribution of time spent in building upgrading. It can also be used as a supplement to funnel analysis.

LTV It is an analytical model that analyzes the business value of users, which can analyze the per capita value contributed by the user group visited on a specific date over a certain period of time.

That’s a brief overview of what behavior analysis covers, and you can click through to each chapter to learn more. If you are interested in user analytics, check out user analytics.

1.Events

1.1 The significance of Events

The behavior data generated by the USER in the APP or game is called the EVENT, and the event data records the time, type and detailed information of the user’s various behaviors. The model that filters or groups these event data and finally calculates aggregate metrics is called the event analysis model.

Through event analysis, we can calculate aggregate indicators of specific behaviors generated by users in a period of time, understand the user participation of each behavior and the development trend of indicators, so as to obtain macro control of products and optimize decisions.

In the case of consumption events, for example, the following problems can be solved:

1. What is the daily income flow trend in the last month?
2. The daily income flow and distribution of each channel in the last month?
3. For users from Shanghai, what is the trend of the amount paid per capita in recent month?
4. How do users from Shanghai and Beijing compare the total number of times they paid for membership cards in the last month?

1.2 Location of Events

Select “Analysis” from “Events” in the top navigation bar to enter the Events model:

1.3 Use of Events

1.3.1 The selection of analysis indicators

The analysis indicator is “event” + “attribute” + “calculation method” or directly “Event” + “calculation method”.
For example: the number of users triggered by any event, the number of per capita purchases of goods, the average VIP level of users who buy goods, the number of users who buy goods using WIFI is true, and the number of de-replications of users who buy goods in provinces.

1.3.1.1 Additive formula

If you need to analyze some of the four operations between indicators, you can do it by “adding formulas”. For example, the ARPU value (the amount paid per active user) can be analyzed with the “order event.” Total order amount. Total “/“ Login game. Trigger user number”

1.3.1.2 Add multiple metrics

You can analyze multiple analysis indicators at the same time by “Add Indicators” to achieve some comparison and display requirements, as shown in the following figure:

1.3.2 Selection of grouping items

Group items can be grouped by user attributes, event attributes, user clusters, or user labels.
For example, if the analysis index is “total order amount of order event” and the group Angle is “channel”, the total order amount of each channel can be analyzed.

1.3.2.1 Set the conditions for the specified group

“Designated group” can be set by combining screening conditions. For example, in the above case, if “District service id” is equal to “1/2” in the screening conditions, “the sum of the respective paid amounts of 1 and 2 district service id recharge events” can be analyzed, as shown in the figure below:

1.3.2.2 Set multiple grouping

User groups can be divided into more detailed dimensions by selecting multiple grouping conditions. For example, in the above case, if the grouping item is selected as “District Service id” and “channel”, the “sum of payment amount of each paid event” of each group such as “1, Google Play”, “2, app store” can be analyzed.

1.3.2.3 Set packet mode

For numeric and temporal attributes, you can set the grouping mode when they are used as grouping items.

【 Numerical attribute 】

【 Time-type attribute 】

1.3.2.4 Setting Event Breakdown

Group Item is set to group all events in an analysis object. Event Breakdown resolves the need to collect statistics on only one event in an analysis object.
In event analysis, if you add two or more analysis objects or select Add Formula, you can set Event Breakdown on the right of the group item:

Click “Event Breakdown” to select an event (except any event) in the analysis indicator to be split and calculated according to event attributes.
As shown in the figure below, the participating activities are divided and calculated according to the activity type, and the analysis index 1 is to calculate the participation rate of different activity types.

After selecting the event split, you can click the split icon in the analysis indicator to set the status of the event participating in the split.
In the figure below, indicators in the formula participate in the splitting, while non-formula indicators do not participate in the splitting, that is, the total number of participation paid for calculation and the participation rate of different payments.

1.3.3 Selection of filtering criteria

By analyzing indicators, you can select events or users that match certain characteristics for analysis
This parameter is displayed when the filter criteria are greater than or equal to two. The default value is “and”. You can click to change it to “or” to adapt to different application scenarios.

1.3.4 Report setup, display, download and save

When the Analysis Angle Setting area is fully selected, click Start Calculation to display the chart result in the Chart Display Area.

1.3.4.1 Shows the filter area Settings

【 Analysis particle size 】
The default analysis time granularity is by day. The supported time granularity can be by day, minute, hour, week, month, and total
【 Analysis period 】
The default analysis period option is “last 8 days”, after clicking, the date selector will pop up, you can select any time
Shortcut options: The left side provides some shortcut options for easy selection, including: yesterday, today, last week, this week, last month, this month, last 7 days, last 7 days, last 30 days, last 30 days, from date to yesterday, from date to date
User-defined: To meet the requirements of user-defined time range, you can customize the start time and end time on the right. Static time and dynamic time are supported
Static time: Select a specific date that does not change with the current date, such as 2022-08-23
· Dynamic time: Select the relative time range of “Current Date”, which changes with the current date.

Note: When the end time is Static time, only static time can be used for the start time.

【 Display analysis indicators and display groups 】
Display analysis objects: All analysis objects are displayed by default. A maximum of 10 analysis objects are displayed. You can modify the objects to be analyzed in the check box.
Show groups: You can modify the groups you want to analyze in the check box.

1.3.4.2 Presentation of report contents

【 Chart area 】
The chart title shows the display name information of the analysis object.
The legend section displays legends according to the total number of legends composed of analysis objects and groups.
Presentation data will be determined and presented according to analysis period and analysis granularity, analysis object and display grouping, and chart style.
Display of line trend chart:

【 Form area Display 】
You can filter the group items, for example, you can search “Beijing” when you group by province.
If you need to swap rows and columns, you can switch By Time, by Event, and By Group if there are group items

1.3.4.3 Query and download event details

Select the total number of events analyzed, and the user can enter the event details by clicking on the link in the table

1.3.4.4 Retention of reports

When saving, you need to enter [Report name] and [Report Remarks] to facilitate daily viewing. After saving the report, you can edit and modify the report in Report Management.

2. Retention

2.1 The significance of Retention

Retention analysis mainly analyzes the overall participation degree and activity degree of users, and examines the number and proportion of users who will return to a certain initial behavior.
Through the retention analysis model, it is possible to analyze the return visits of users who complete an initial event within a period of time, so that the impact of the event can be macro-controlled and the decision can be optimized.
For example, from user activity to consumption, the following problems can be solved:

1. Within 1 month after users use the product, the number and proportion of consumption transformation?
2. The number and proportion of users who complete consumption within 1 month after using the product, and the consumption amount reaches 200 yuan?
3. The number and proportion of Shanghai users who have converted their consumption within 1 month after using the product?
4. The number and proportion of first-tier city users who completed consumption within 1 month after using the product, and the consumption amount reached 200 yuan?
5. What is the 30-day LTV status of full users

2.2 Location of Retention

Select “Retention Analytics” from “Analytics” in the top navigation bar to access the Retention Analytics model.

2.3 Page overview of Retention

Retention analysis is divided into three parts: indicator setting area, display setting area and display chart area. You can determine the content of the analysis by setting the analysis angle. And by selecting the display filter option, the required analysis table or graph is drawn in the chart area

2.4 Usage Scenarios for Retention

2.4.1 Setting of analytical indicators

2.4.1.1 Initial event selection

Users who have had an initial event behaviour will be used as a data source for the analysis sample and will play a role in the overall analysis. The checkboxes allow you to select a “meta-event” or “any event”. This means that users who have done this event on a specific date. Users who did any event on a specific date can be interpreted as “they were active during that time”.

2.4.1.2 Selection of screening conditions

[Event Filtering]
Filtering conditions are mainly designed to accurately view the usage of specific groups of people, such as analysing the retention of male users in Shanghai. You can filter both initial and return events. Through the “filter conditions”, you can select events or users that meet certain characteristics and then analyse them.

[All events meet]

2.4.1.3 Selection of return events

Returning users are users who have completed a return event after a specific period of time for the initial event user. A “meta-event” or “any event” can be selected in the checkbox. That is, the user who has done the return event on a specific date. Users who have done any event on a specific date can be interpreted as “they have been active during that time”.

2.4.1.4 Selection of Calculate another metric

[Meaning of Calculate another metric]
While analysing and filtering the retained users, we may do some more in-depth analysis on the retained users.
For example, to independently analyse the active, paid, LTV and other data of the retained users, we can use the simultaneous display function to deeply analyse some specific indicators.

[Entry for Calculate another metric]
When clicking on “Calculate another metric”, the content to be analysed will appear, and you can select the required events and analysis angle through the drop-down box.

[Analytical perspectives for Calculate another metric]
Similar to event analysis, Calculate another metric of metrics can be “Event” + “Attribute” + “Calculation Method” or directly “Event” + “Calculation Method”.

2.4.1.5 Group by

[Setting of Group by]
Event attributes and user attributes can be selected as grouping items.

[Grouping Presentation]
Under the data table style, click the “+” button to view the group details in the pop-up window.

2.4.2 Setting of Display Setup Area

2.4.2.1 Period of analysis, duration of analysis and indicators

• You can restrict the data to be analysed for a specific period of time through the “Analysis Period”.
• You can switch the performance of retention and churn under different volumes through “Analysis Period”.
• You can select the angle of analysis through “Indicator”, such as “Retention, Churn”.

2.4.3 Chart presentation and data download

2.4.3.1 No increase in simultaneous display of display content

2.4.3.2 Increase the number of displays that can be shown simultaneously

2.4.3.3 Only display content that is displayed at the same time

2.4.3.4 Text alert

Mouse over the chart box and a text prompt will be displayed.

2.4.3.5 Data Download

The style of the downloaded table is the same as when the chart is styled as a table, in csv format.

3. Funnel

3.1 The significance of Funnel

Funnel analysis is a kind of analysis model to analyse the conversion of users in the specified steps of the behavioural flow, which can help analysts quickly grasp the conversion of the product in each step of the process over a period of time, so as to achieve the purpose of checking the gaps and filling the gaps, and optimize the conversion process.
The following business scenarios can be solved through funnel analysis:

1. Analyse the conversion of users who log in and participate in daily activities.
2. Analyse the conversion of users who register, log in, complete the novice guide, participate in activities, consume props, and pay.
3. Analysing the difference in the conversion of payment by platforms

3.2 Location of Funnel

Select “Funnel” from “Analytics” in the top navigation bar to access the Funnel Analytics model.

3.3 Overview of the Funnel Analysis page

Funnel analysis model consists of “indicator setting area, display filter area, display chart area, table details area”.

1. In the “Indicator Settings Area”, set the “Analysis Steps, Grouping Methods, Analysis Window Period”.
2. In the “Display Filter Area”, set “Analysis Time Period, Analysis Steps, Group Settings”.
3. In the “Show Chart Area”, you can mouse over a grouped object to view the details and see the corresponding “single-step conversion rate” of each step, and set the “chart style”.
4. In the “table details area”, you can view all the funnel data information of the subgroups.

3.4 Scenarios for the use of Funnel

3.4.1 Setting of the indicator setting area

3.4.1.1 Setting of the indicator setting area

(1) “Funnel Steps” are two steps, respectively, two events in metadata management, you can switch to select an event (you can not select any event) or add steps.
(2) “Global Filter” can filter the event attributes or user attributes common to all steps.
(3) “Grouping items” is empty by default, i.e., no grouping, you can select one of the event attributes, user attributes, user subgroups and user labels in step 1 to add.
(4) “Analysis Window Period” is 1 day by default, you can select the same day in the drop-down box or a custom window period by day, by hour, by minute.

3.4.1.2 Adding, deleting and filtering conditions in the analysis step

1. You can insert a new step by clicking the “Add Step Indicator” button or by clicking the “+” button of a step, and you can add up to 30 steps.
2. You can delete a step by clicking on the “Delete button” of the step, and the serial number of the other steps will be changed after the step is deleted. In addition, the “Delete Button” is not displayed when there are only 2 steps.
3. You can add filter conditions by clicking on “Add Condition” of a step, the logic of adding, deleting and changing filter conditions is the same as that in event analysis.

3.4.1.3 Definition of the analysis window period

The analysis window period, i.e., the total analysis timeframe counting from the start of the triggering step 1 using the user who triggered step 1 as a sample, is fixed in length for each user, but the start and end funnel absolute dates may be different due to the point in time at which step 1 was triggered.

3.4.2 Setting of Screening display area

3.4.2.1 Default Initial Page

3.4.2.2 Definition of the analysis period

The analysis period determines the selection interval for “Trigger time for step 1”, the drop-down box is the same as for event analysis.

3.4.2.3 Operation of step checkboxes

The Steps checkbox selects all steps by default, or you can select steps from M to N in steps.

3.4.2.4 Grouping Setup

Chart settings differ from event analysis in that the overall always appears as the first item in the grouping
When more values are set for grouping, 6 groups are selected by default

3.4.3 Setting of Chart display area

3.4.3.1 Cumulative conversion funnel content

(1) Using the first of the selected steps as 100%, calculate the number of conversions after each step and give the ratio
(2) In the display of each step, the conversion rate of the previous step is indicated by the background fill colour.
(3) When the mouse moves into a grouped object, the grouped object is highlighted, showing the corresponding conversion rate, the number of people converted and the conversion rate of a single step in the middle of each step (the colour is the same as that of the group into which it is moved).

3.4.3.2 Elements of a conversion trend chart

Calculate the daily conversion rate from the beginning of the selected step until the last selected step, using the first step of the selected step as 100%.

3.4.4 Setting of Table display area

3.4.4.1 Tabular content of conversion charts

Show all subgroups (both overall and all possible groups, up to 1000 items), showing the cumulative conversions to each step and colour-coded (similar to retention tables)

3.4.4.2 Tabular content of trend charts

4. Distribution

4.1 The significance of Distribution

The distribution analytics model allows you to divide intervals based on the total number of completions, days, or attribute values aggregated for each user (the subject of the analytics) and view the number and percentage of users in different intervals. The following are common analysis scenarios
Event frequency (number of times): divide the interval by the number of times a user participates in battles per day, and view the number of users with different number of battles per day
User stickiness (days): by the number of days users logged in over the past seven days, differentiate between users with different stickiness for subsequent analysis
Attribute Value: Split users into large, medium and small Rs according to the cumulative payment amount of recharges over a period of time, and view the proportion of users of different levels.

4.2 Location of Distribution

Select “Distributional Analytics” from “Behavioural Analytics” in the top navigation bar to access the Distributional Analytics model.

4.3 Overview of the Distribution Analysis page

Like other models, the distribution analysis consists of three parts: “Indicator Setting Area”, “Filter Display Area” and “Chart Display Area”.

4.3.1 Indicator setting area

4.3.1.1 Setting up of user participation events

Similar to Event Analysis, Distribution Analysis can choose (“Event” + “Attribute” + “Calculation Method”) or directly (“Event” + “Calculation Method”) when setting up user participation in an event, but there is a difference between the supported calculation methods and Event Analysis.

4.3.1.2 Setting of distribution intervals

Clicking on the cogwheel icon opens the interval settings, by default the “Default Interval” is selected, and the logic of all the options is exactly the same as the interval settings of the numeric type attribute in the grouping item.

4.3.1.3 Setting of filter conditions

[Setting of filtering conditions]
You can filter the event attributes or user attributes or user subgroups under the event, and the filter conditions support and or logic, consistent with the event analysis.

4.3.1.4 Selection of Calculate another metric

Users based on a certain distribution interval can continue to be analysed in depth by using the Show Simultaneously option.
Clicking on the “Show Simultaneously” button, a new option will appear, named “Show User Engagement by Zone Simultaneously”, which can be selected as (“Event” + “Attribute” + “Calculation Method”) or directly (“Event” + “Calculation Method”), and the options are the same as those in the event analysis model.

4.3.1.5 Group by

Grouping items can be selected from event attributes or user attributes of the event

4.3.2 Screening display area

4.3.2.1 Initial Analysis Page and Options

There is a choice of analysis periods, analysis granularity, grouping checkboxes, and the ability to export tables or download the full amount of data. If the Simultaneous Display option is set in the Indicator Settings area, the option “Only see simultaneous data” is displayed in the Filter Display area.

4.3.2.2 Default settings and selection range for each option

Analysis Period： The time range interval for analysing the data will be determined, with “Last 8 days (dynamic time)” selected by default.
Analysed Particle size： Default is “By day”; you can choose “By day”, “By week”, “By month”, or “Total”. When “Weekly” or “Monthly” is selected, the analysis period will be supplemented according to the actual granularity, and the supplementation method is the same as that of the event analysis model.
View [Also Show] Only： Unchecked by default, when checked, all charts in the chart display area will only be plotted according to the indicators displayed at the same time.

4.3.3 Chart display area

The “calculate another metric” indicator is not set in the left indicator setting area.
There is a “calculate another metric” button in the left indicator setting area, but the “View [Also Show] Only” option is not ticked in the filtering display area.

4.3.3.1 Tables

The main structure of the table
-The first row is the title of the table, from left to right are the time of the event, all users (subject of analysis), interval 1, interval 2 ……
-The contents of the table from top to bottom are: the number of users in the interval (the subject of analysis), the percentage, and the indicators displayed at the same time.
-When the mouse is moved to the table, the data logic corresponding to the cell will be displayed.

Event Time
-The interval of event occurrence time is determined by the analysis period.
-Split and complete the time period according to the granularity of analysis, such as “2021-09-06 week”, “2021-09 month”.

Presentation logic for grouping
-Grouping items can be put away or expanded. When expanding the grouped items, you can expand the grouped items under multiple dates.
-Each user’s group attributes are counted according to the attributes at the time of the actual behaviour.
-Record a total of 1000 possible “date+group” data, with the option to “download full data in page format” for complete data.
-Supports drill down to the user list, which allows you to underline the user list and further view the sequence of user behaviours for a single user.

5. Flows

5.1 The significance of Flows

Path analysis is an exploratory model that analyses the sequence of behaviours, behavioural preferences, critical nodes, and conversion efficiency.
Path analysis records the order of user visits during each session, and then integrates it to obtain a Sankey diagram of user behaviour paths, intuitively viewing the inflow and outflow of behaviour before and after each key node.
It can intuitively grasp the expansion route of user behaviour for optimizing the content of nodes and improving the overall conversion efficiency. Using the path analysis model, you can quickly understand the primary and secondary factors affecting the conversion, so as to improve the product with a purpose.

5.2 Location of Flows

Accessible from the “Flows” page under the “Analysis” module.

5.3 Overview of the Flows Analysis page

It consists of four parts: analysis angle setting area, display filter area, display chart area, and node detail information.

5.4 Usage Scenarios for Flows

5.4.1 Definition of terms

Analyse Event Groups
Default 10 events and maximum 30 events can be selected to participate in the route analysis. Unselected events are invalid events and do not participate in the calculation and display of path analysis.
[Objects for analysis].
Selects an item in the analysis event group as the basis for analysis, and you can specify it as the initial or end event to be analysed. By default, the first starting event is selected.
Session interval
If the trigger interval between two consecutive valid events (belonging to the analysis event group) is within the set time, they are considered to belong to the same session.
[Path].
The path is expanded forward or backward (determined by the beginning or end of the analysis) based on the object to be analysed. The paths will be independent of each other and fully exhaustive.
[Number of displays].
When a node on the path is in the following cases, it will be merged into the “more” module in the chart.
Note: A node will be merged into “More” if the number of nodes in that node is more than the 7th highest in its class.

5.4.2 Analysing the conditions of the angle-setting zone

5.4.2.1 Events analysed with participation

The first 10 items are selected by default, up to 30 items can be selected, support the search function

5.4.2.2 Setting Group Split Criteria for Events

An event split option can be added for the selected group of Events Involved in Analysis. At this point:

1. Any event in the group can be split according to an event attribute of the event.
2. Each event can only be split by one attribute, i.e. the selected event cannot be selected again.
3. Each event in the group can be split.
4. Numeric, temporal or list attributes can be set to intervals, and the attribute interval groups share the configuration with the interval groups of other models.

5.4.2.3 Selection of events for analysis

Selection of analyses and setting of filter conditions

1. The mode of analysis can be selected from the initial event or the end event.
2. Filter conditions can be added to the event attributes of the event.
3. Filter conditions can be analysed from the perspective of the data type.

5.4.2.4 Session interval settings

The meaning of session interval is: The maximum time between neighbouring events when selected.
Modify this value to adjust whether the neighbouring events are in the same session.
The default setting is “30 Minutes”, and the unit can be “Seconds” (160), “Minutes” (160), “Hours “(1~24)

5.4.3 Setting of Screening display area

Default selection is the last 8 days

5.4.4 Setting of Chart display area

5.4.4.1 Lost or starting displays

1. when analysing the start event, there is no lower level node at a step, then that node will appear to be lost
2. when analysing the end event, a step has no upper level nodes, then the node will appear to start

5.4.4.2 Components of “more”

When the node is “more”, it is equivalent to the set of general nodes, and the selected lines are the set of lines, values and flows of all the constituent nodes.

5.4.4.3 Horizontal scrollbar with left and right drag

Up to 10 levels can be displayed, with the default display showing the initial event or the end event (depending on the choice).
More information can be displayed in the screen by dragging the horizontal scrollbar.

5.4.4.4 Node Path

When clicking on a node, you can highlight the relevant path or display the node information details of that node.

When viewing node details, clicking on the number of users information will display a list of eligible users. And the related user behaviour sequence can be queried

6. Interval

6.1 The significance of Interval

Interval analysis can be used to analyse the time interval between events to understand how often a core user behaviour occurs, or as a complement to funnel analysis, where you can get a more detailed view of conversion durations in addition to analysing conversion rates between steps.

In the interval analysis, you can get answers to the following questions:

• How long does it take, on average, for a user to sign up and pay for an icebreaker?
• Is there an anomaly in the elapsed time from opening the app to actually getting to it, and is there a problem with the installation package from one of the channels?
• How long does a user need to reside in the first chapter before moving on to subsequent chapters?

6.2 Location of Interval

Select “Interval” from “Analysis” in the top navigation bar to access the Interval Analysis model.

6.3 Overview of the Interval Analysis page

In interval analysis, you need to select the analysis subject and the start event and end event. Only when the same analysis subject completes the start event and the end event in sequence will it be regarded as one interval.

6.4 View interval statistics

Interval analysis offers two chart types that you can choose from depending on the purpose of the analysis

• Box-and-whisker plot
  If you wish to see aggregated statistics for event intervals, such as maximum, upper quartile, median, lower quartile vs. minimum, you can select the box-and-whisker plot display. If you select a time granularity other than aggregate, the box-and-whisker plot will show the variation with date.

• Histogram
  If you wish to see the statistics of the distribution of event intervals, you can choose the histogram display.

7. LTV

7.1 Significance of LTV analysis

LTV analytics is an analytical model for analysing the business value of users, which analyses the per capita value contributed by a group of users visiting on a specific date for a certain length of time.

四、Users

Based on Users, you can create segments or tags to help with analysis in addition to using user attributes for analysis.

Attribute AnalysisIt is a model dedicated to analysing the statistics and distribution of user attributes, through which you can quickly grasp the user profile of a specific user group, providing help for refined operation, such as comparing the cumulative paid amount of the average value of users in different channels, or viewing the number of users by the user’s current level and the VIP level of the split to view the number of users.

CohortsIt supports the formation of user groups of people with common characteristics, macroscopic understanding of the group characteristics of various groups of people and microscopic insight into specific user information, and facilitates the segmentation and analysis of user samples in a variety of models, such as finding users who have topped up their account in the past 30 days or those who have not had any events in the past 7 days (churned users).

TagsIt supports calculating user data in a specific way, and the tagged value can be regarded as a special user attribute, such as the number of login days in the last week or the recharge amount since the opening of the service. The tag also supports daily automatic backup, which is convenient to view the historical trend of the number of users corresponding to the tag value, such as the change of the number of big Rs in the game.

user searchIt is easy to find specific users using user attributes as filtering conditions, such as customers searching by account ID, or finding a group of users for subsequent marketing recalls by registration time.

User Behaviour SequenceIt is possible to dig deeper into the behaviour of individual users to understand user behaviour at the micro level and combine it with better analysis at the macro level (analytical models).

1. Attribute Analysis

1.1 The significance of attribute analysis

Attribute analysis is a model that specialises in analysing the statistics and distribution of user attributes. The model categorises users according to their attributes and can simultaneously view the statistics and distribution of users with different grouping values. For example, it can view the distribution of users in each province, the age distribution of users, and so on. The user profile of the overall user group can be quickly depicted. Multi-angle, omni-directional grasp of the characteristics of the designated user groups, macro grasp of the composition and preferences of the overall user, so as to provide the basis for the refined operation.

1.2 Location of attribute analysis

Accessible from the Attribute Analysis page under the Users module.

1.3 Page overview of Attribute Analysis

By “analysis angle setting area”, “display filter area”, “display chart area”, “table details area”. Four parts

1.4 Scenarios for the use of attribute analysis

1.4.1 Conditions for analysing attributes

The content that can be analysed is the number of users, user attributes of a certain analysis perspective
Note: If you select the number of users as the analysis attribute, you can click on the user list in the table area.

1.4.2 Group by

The default is no grouping, in which case the entire user population is analysed directly without distinguishing between them.

2. Cohorts

2.1 The significance of Cohorts

By forming user groups from people with common characteristics, it is possible to understand the group characteristics of various groups of people and gain micro insight into specific user information, and to facilitate the segmentation of user samples for analysis in various models.

• Conditional subgroups: user subgroups obtained by filtering specific behavioural conditions or user attributes.
• ID Segmentation: Generate user segments by uploading ID files based on the content of the uploaded fields.
• Result Segmentation: User segments obtained by using the results of the analysis model.
• SQL segmentation: user segments are created by querying the results of SQL statements.

2.2 Location of Cohorts

Users who have the privilege to view user groups are accessed from “User Analysis - User Groups”.

2.3 Page overview of user segments

2.4 User Segmentation Usage Scenarios

2.4.1 List of subgroup base information

The list of subgroups will display subgroup name, display name, subgroup note, creator, subgroup type, data update, creation time, update time, number of users, and operation column.
Click “User Group Number” to view the list of users (analysis subject) corresponding to the subgroup, and “Operation” column can be set up differently according to different subgroup types:

cluster types	action items
Conditional Grouping	Edit/Download/Refresh/Delete
ID Grouping	Edit/Download/Refresh/Delete
SQL Cluster	Edit/Download/Refresh/Delete
Result Subgroups	Edit/Download/Remove

Note: Members with relevant permissions within the project can view user groups created by other users and the corresponding user lists, but can only edit or delete user groups they have created.

2.4.2 Conditional Cohort

2.4.2.1 New conditional Cohort

Click “+New Subgroup” at the top right of the subgroup page, and select “Conditional Cohort” (note that there is an upper limit for user groups, and they cannot be created when the upper limit is reached).

• Grouping conditions
  There are three types of conditions that can be added: events that have been done, events that have not been done, or user attributes that are satisfied, and you can add any one of them.

Conditional Cohort can choose between two update methods, “manual update” and “automatic update”:

• Manual update” means that after the first calculation, the system will not update the user group automatically, users need to update manually.
• “Automatic update” means that after the Cohort is created, the system will calculate and update the Cohort data at the specified time of the day, using the previous day as the base date.

2.4.2.2 Viewing, editing and deleting conditional Cohort

• You can enter the group settings page to check the conditions of the group.
• General users can only edit and update their own usergroups.
• General users can only delete the subgroups that they have created, the delete button will not be displayed for the groups that they have not created and they cannot be deleted.

2.4.3 ID Cohort

2.4.3.1 New ID Cohort

Click “New Cohort” at the top right of the group page and select “ID Cohort”.

Go to the creation page and upload the file

• The format of the uploaded file should be as follows: one ID field per line in the ID file, recorded in UTF-8 encoded CSV text format.
• ID upload type label, can not set the “update method” and “backup method”.
  You can select a user attribute (numeric or text type) as the associated field to judge the user, and any user ID that matches the value of the selected user attribute at the time of upload (the first column) will be assigned the corresponding tag value.

2.4.3.2 Viewing, editing and deleting ID Cohort

• You can enter the subgroup settings page to check the conditions of subgroups.
• General users can only edit and download their own ID Cohort.
• General users can only delete their own ID Cohort, user groups not created by them will not show the delete button and cannot be deleted.

2.4.4 Result Cohort

2.4.4.1 New Result Cohort

In the event analysis model, if you are counting the number of triggered users, you can create Cohort in the results table by clicking on “Create Result Cohort”.

In the case of the path analysis model, you need to click through to the node details before the Create Resulting Cluster button appears.

For other models, such as retention analysis, you can click on the Create Outcome Cluster button wherever it appears in the table.

• When creating a subgroup, you can set the name of the results Cohort and a note to describe the results Cohort.
  

2.4.4.2 Editing and deletion of Result Cohort

• Result Cohort can not modify the creation rules and update methods, only the name of the Cohort and notes.
• General users can only delete the result Cohort they created, user groups not created by them will not show the delete button and cannot be deleted.

2.4.5 SQL Cohort

2.4.5.1 New SQL Cohort

Click “New Cohort” at the top right of the group page and select “SQL Cohort”.

• Create SQL Rules
  According to the required logic, write the corresponding statement in the SQL statement input box, and the SQL Cohort will be generated based on the SQL rule query result.

• SQL Dictionary
  On the right side of the creation interface, SQL dictionary is provided, which is convenient to view the current available table structure, including event table, user table, dimension table, temporary table, etc., and copy the corresponding field names or parse them directly.

SQL Cohort allows you to choose between two update methods, “Manual Update” and “Automatic Update”:

• Manual update” means that after the first calculation, the system will not automatically update the user group, users need to update manually.
• “Automatic update” will calculate and update the user group based on the base date after 0 o’clock on the server.

2.4.5.2 Viewing, Editing and Deleting SQL Cohort

• You can enter the group settings page to check the conditions of the group.
• General users can only edit and update their own user groups.
• General users can only delete their own SQL Cohort, the delete button will not be displayed for user groups not created by them, and they cannot be deleted.

2.4.6 User Cohort List

Click “User Cohort size” of a subgroup in the subgroup list to enter the user list page, which displays the user list information of the eligible users.

3. Tags

3.1 The significance of Tags

The Tags is a collection of people with a certain set of characteristics, and the tagged value is a collection of “a group of similar people”, which can be easily structured by the label and tagged value.

For example, if the Tags is “Paid User”, there are three groups under the label with label values of “Big R”, “Medium R” and “Small R” respectively. Under the label, there are three groups with label values of “Large R”, “Medium R”, and “Small R”.

Through a certain creation method, the users with similar characteristics are aggregated by characteristic value and then form a label, which is convenient for segmentation analysis in various models.
1、Provide a variety of creation methods, suitable for more in-depth user group drill-down analysis.
2、Multiple similar populations in the same dimension form a label, and each population is treated as a label value to achieve the connection between related populations.
3、Tags can be created for different dates of the historical version (backup), you can see the number of users with different tag values with the date of change
4、Tags can be switched to different analysis subjects, corresponding to different calculation logic and analysis scenarios.

3.2 Location of Tags

Users with Tags viewing privileges enter from “User - Tags” to set up Tags.

3.3 Overview of pages with Tags

Click the “New Tag” button and select a label type.

Complete the entire label creation process

See how different users are tagged

Use of labels in model filtering, grouped item viewing

3.4 Scenarios for the use of user tags

3.4.1 User tags and tagged values

User label is a set of “the same dimension, a number of similar groups of people” collection, label value is “a set of similar groups of people” collection.

For example, if the user label is “Paid User”, there are three groups under the label with label values of “Large R”, “Medium R” and “Small R” respectively. There are three types of groups under the label with label values of “Large R”, “Medium R” and “Small R” respectively.

Users belonging to a tag have unique tag values, and if a tag has a historical version, the tag value of the user may be different for each historical version (each date) of the tag.

3.4.2 Creation of user tags

When you create a new one, you can choose one of the three types of tags: “Conditional tag”, “ID tag”, “SQL tag”, and then enter the specific creation configuration page.

3.4.2.1 Tag Definition

1、Conditional Tag

Filter users for specific behavioural conditions or user attributes and assign tag values.

For example, you can create a “paid user tag” and then classify users into high-consumption users, medium-consumption users, and low-consumption users by customising the conditions. The tag can then be used as a grouping item to analyse all three types of users at the same time.

• The users in all tagged values together form the tagged user
• New tagged values are always added after the last tagged value.

2、ID Tag

Define the tagged user according to the canonical upload ID required by the template and assign the tagged value at the same time, with the option of selecting a user attribute as the associated field to judge the user

• Users with the same value in the second column will form the same tagged value
• If the user does not have a tagged value, then the row will be discarded and an alert will be given in the last step of the results display.
• ID Tag for upload type, cannot set “update method” and “backup method”.

You can select a user attribute as the associated field to judge the user, and any user ID that matches the value of the selected user attribute at the time of upload (the first column) will be assigned the corresponding tag value.

3、SQL Tag

Generating tagged values for analysis subjects based on SQL rules

(1) Tag condition settings at creation

According to the required logic, write the corresponding statement in the SQL statement input box, the query result needs two columns, the first column is the non-repeated analysis subject ID and the second column is the tag value.

(2) SQL Dictionary
On the right side of the creation interface, a SQL dictionary is provided to facilitate viewing the currently available table structures, including event table, user table, dimension table, temporary table, etc., and copying the corresponding field names or directly parsing the

3.4.2.2 Updates and Backups

Only when the tag type is not “ID Upload”, you can set the update method, the default option is “Manual Update”.

• Manual Update: After creation, the label data will be fixed, you can click “Update Data” button to calculate and update the label data manually.
• Automatic update: After creation, the system will calculate and update the tag data at the specified time of each day with the previous day as the base date.

3.4.3 User Tag List Page

• By clicking on the tag name, you can enter the user’s tag details page (data details).
• You can create a tag by clicking the “New Tag” button in the upper right corner.
• Click “Edit Tag” to enter the tag editing page.
• Clicking “Update Data” will recalculate the latest tag version using the current date as the base update date.
• Click “Download” to download user tag data directly.
• Click “Delete” to delete the tag.

3.4.4 User tag detail page

User tag detail page with tag base information, data details

3.4.4.1 Tag base information

Display the basic information of tags, which is basically the same as the content displayed on the user tag list page.

• Click “View Configuration” to open the detail pop-up window, to see the specific conditions of creation
• Click the “More” button next to the tag name, you can operate, the same as the user tag list page.

4. User Search

4.1 The significance of user searches

User search is set in the navigation bar, so that users can search for matching users based on specified conditions on any page.

4.2 Location of user searches

It can be accessed from the search ICON in the navigation bar

4.3 Overview of pages searched by users

4.4 Usage scenarios for user search

4.4.1 Search scope and query matching method

1. All “User Attributes” and “Account ID” sections can be searched.
2. When filtering, only results that match exactly will be searched.
3. You can enter the user behaviour sequence of the user by ID.

4.4.2 View a sequence of user behaviours for a single user

Clicking on a specific user takes you to that user’s specific behavioural sequence. This makes it easier to count user behaviour and track that user’s specific behavioural trajectory.

5. User Behaviour Sequence

5.1 The significance of user behaviour sequences

Allows microscopic observation of the specific behaviour of users who match certain characteristics. It is convenient to filter the attributes and do further analysis of the user’s specific behavioural distribution, behavioural paths, and behavioural characteristics preferences in a specific time zone. Thus achieving the purpose of improving efficiency and optimising decision-making.

5.2 Location of user behaviour sequences

To view the sequence of user behaviours, you first need to click into the list of users in the calculation results of the model, and then by clicking on the underlined IDs in the list of users, you can enter the sequence of user behaviours. The following are the main entry points:
The “User List” page of the event analysis model where the analysis dimension is “Number of Triggered Users”.
The “User List” page in the Retention Analytics model for “Number of Retainers, Number of Churners”.
The “User List” page for “Conversions, Churn” in the Funnel Analytics model.
The “Users” page to see the number of people in each segment in the distribution model.
The User Details page to view node information in the path analysis model.

5.3 Page overview of user behaviour sequences

5.4 Scenarios for the use of user behaviour sequences

5.4.1 Components of a user behaviour sequence page

5.4.1.1 Default Initial Page

Includes user attributes area, behavioural statistics area, behavioural sequence details area

1. User attribute area: view the user attributes of the current user.
2. Behaviour statistics area: you can view the frequency of the current user’s behaviour according to the time granularity.
3. Behavioural Sequence Details: view specific behavioural sequences and event attributes of the current user.

5.4.2 Contents and operations of the User Properties area

5.4.2.1 Filtering User Attributes

The first time you enter the User Behaviour Sequence for an item, all basic user attributes are selected in full by default, including preset and custom attributes.
The drop-down box allows you to filter the user attributes that need to be displayed, and after filtering, the display will be changed.

5.4.3 Content and operation of the behavioural statistics area

5.4.3.1 Initial page (bar daily trend)

By default, you can view a trend graph of the number of daily behavioural events for the last 7 days.

5.4.3.2 Screening behavioural events

The first time you enter a user behaviour sequence for a project, all behavioural events are selected by default.
The drop-down box allows you to filter the behavioural events that need to be analysed. After filtering, the display will change.

5.4.3.3 Settings and Effects of Time Selection

By default, the “Last 7 days” time selection is selected and the granularity of the timeline is “By day”. This displays a graph of the trend of the number of behaviours for the last 7 days.
You can click to select different dates in the bar chart. In this case, the dates under Behavioural Sequence Details are anchored to the corresponding days.
The granularity of the timeline is “By Day” and “By Hour”.

5.4.3.4 Calculated content of trend graphs

The charts will only count the filtered events and the trend charts will show the behavioural statistics at each time interval.

5.4.4 Content and operation of behavioural sequence details

5.4.4.1 Initial Page and Event Expansion Page

Initially “the last date for which data are available in the Date checkbox”, and is equivalent to selecting this date.
If you select “By day” as the analysis interval, you can only see the data under the selected day.
If you select “Hourly” as the analysis interval, you can only see the data for each hour of the day under the selected day.

5.4.4.2 Expand Event Details to see event properties

By clicking on an event, all event properties under that event will be expanded

五、Data

1. The significance of Data

Data management is mainly used to manage the presentation of all reported data, including modifying the display status of all events, event attributes and user attributes in the system; and adjusting the notes of events, event attributes and user attributes.

2. Location of Data

In the top navigation bar, “Events”, “Event Properties” and “User Properties” in “Data Management” are the categories of data management. “are the categories of Data.

3. Overview of the Data page

Data management is divided into three parts: event management, event attribute management, and user attribute management.
Each part can be searched individually, and after the setup is complete, the events, event attributes, and user attributes will be displayed as new in each analysis model.

4. Data Usage Scenarios

4.1 Query Events

Events can be retrieved by filtering the specified conditions, all events are displayed by default, and the display status can be modified after clicking Display.

4.2 Query the event properties under a specific event

All event attributes under a specific event can be queried.

4.3 Querying and Editing Event Properties

Same as the query editing of event management, you can retrieve event attributes by filtering the specified conditions, and all event attributes are displayed by default.

4.4 Querying and Editing User Attributes

Same as query editing in event management, user attributes can be retrieved by filtering the specified conditions, and all user attributes are displayed by default.

5. Dimension Table Properties

5.1 Meaning of Dimension Table Attributes

For event attributes and user attributes that have already been reported, the originally uploaded data can be mapped to another display or calculated value by uploading a dimension table, so that the attribute values at the time of the initial burial are not the same as the display values, and so that the data can be processed at a later stage to improve flexibility. Dimension table attributes can be used in filtering conditions and grouping items in the analytics model.

5.2 Scenarios for using dimension tables

Dimension Table Attribute: During data collection, channels may be recorded by channel number, for example, A01 for Xiaomi channel (Xiaomi App Store), A02 for Baidu channel (Baidu Mobile Assistant). If you want to use Chinese in the analysis, you can create a dimension table attribute for the “Channel” attribute and set the Chinese name corresponding to the channel number in the dimension table. In the dimension table, you can use the dimension table attribute “Channel Name” with the corresponding Chinese name in the calculation.

5.3 Concept of Dimension Table Attributes

5.3.1 Basic Overview of Attribute Types

• Custom Attributes: Attributes that are customised to be uploaded based on business requirements, such as paid amount, etc.
• Dimension table attributes: newly generated attributes by uploading dimension tables based on existing attributes, which may disappear due to overwriting
  Note: There are five data types: text, numeric, time, boolean, and array.

5.3.2 Creating Dimension Table Properties

5.3.2.1 Create an Entrance

You can create dimension table attributes in Event Attribute Management and User Attribute Management in Data Management.

5.3.2.2 Upload dimension table

1、Please select “Upload” dimension table attribute on the original attribute field where you need to add dimension table.
2、The specific rules of dimension table file

• The first line of content will be the field name of the dimension table attribute, which should start with English, and be composed of English, numbers or “_”.
• The content of the first column will be associated with the original attribute, and the values taken need to correspond to the original attribute values and ensure that the values are unique. If duplicates are encountered, the first one will prevail and subsequent duplicates will be discarded.
  3、Save the csv file in UTF-8 encoding format.
  4、Add the dimension table file to the upload location.
  5、Fill in the display name and field type.
  6、System parsing situation prompts
  7、After finishing, you can find the newly added dimension table attributes in the list.
  8、You can use the dimension table attributes in the model for analysis.
  

5.3.3 Use of Dimension Table Attributes

5.3.3.1 Managing Dimension Table Properties

Dimension table attributes can be managed on the Event Attribute Management or User Attribute Management page.

5.3.3.2 Download & Delete

Dimension table properties can be deleted by selecting the “Delete” option from the popup that appears when you click on the button.