Preface #

In the "Beginner's Guide for Business Users", we have already learned about the main modules of Intelligent Data Insight and their relationships. Next, we will continue to explore some advanced features to help everyone use the platform more effectively.

1. Visual query #

1.1 Field configuration #

1.1.1 Add field - How to use existing fields to create new fields

• For most dataset field requirements, you can consult the data warehouse team. However, if you only need a temporary field for your own analysis, you can also try creating a field yourself. Creating a field requires using expressions, which may be challenging for those who are not familiar with SQL. If you have time, you can refer to this document: Add or Modify Fields.
  

1.1.2 Parameter Management - Define a value that can be flexibly modified for expression calculation

• A parameter is a variable that can be flexibly adjusted. This variable can be used in calculated fields to implement dynamic logic for a single calculated field or even an entire chart. When the parameter takes different values, the field expression will correspond to different calculation logic.

1.2 Chart configuration #

1.2.1 Configure Field Grouping - Grouping metrics in a chart

• For dimension fields, when there is a lot of information and it can be displayed in groups, you can create groups to enable collapsible display of information.
• For numeric metric fields, creating groups allows you to group by the value of an expression field, forming a new field.

Common application scenarios:

• Merging dimension items of text fields: Divide the field values in the "Country" field into "Europe", "Asia", and so on, generate a new group field "Continent", and then use "Continent" as a dimension for analysis.
• Custom data buckets: Divide the sales price into data buckets, such as 0-100, 100-300, 300-500, and observe the performance of data in each range.

1.2.2 Dynamic Dimensions and Dynamic Metrics - Quickly switch dimensions and metrics in a chart

• Add dynamic dimension/dynamic metric functionality to enable simpler and faster switching between multiple dimensions or metrics, meeting the needs of different users for various dimensions or metrics.

1.2.3 Drill Up and Drill Down - Drill down to other dimensions for analysis without modifying the chart

• Drill up and drill down refers to creating a hierarchical structure based on fields (for example: province -> city -> channel), allowing users to quickly drill up or down between levels during visual analysis by clicking.

1.2.4 Conditional formatting - Add solid color, icon sets, color scales, and data bars to charts

1.2.5 Template link - Pass dimension values into links for web page redirection

1.2.6 Data Annotation - Add text annotations to charts

• The visual query module provides a data annotation feature, which supports annotating a specific data point or a specific time period in a chart.
• Use cases: When a data anomaly occurs at a certain data point or during a certain time period for some reason, users can add data annotations at the corresponding position to inform other users of the reason for the data anomaly. This reduces the need for verbal or written communication and improves readability and communication efficiency.

1.3 Filter configuration #

1.3.1 Combination Filter - Perform complex AND/OR condition combination filtering

• The product supports multi-level OR/AND filtering, enabling filtering under complex filter conditions.

1.3.2 Subquery- Use the query results of other charts as filters for the current query

• Dimension filtering supports the subquery feature, allowing users to use the query results of other charts as filters for the current chart.
• Caution:

Charts containing the following conditions cannot be used as subquery charts:

1. A chart cannot reference itself
2. The included subquery charts must also have read or higher permissions for the corresponding dataset
3. Cannot contain dynamic dimensions/metrics
4. Cannot use associated datasets
5. Cannot have parameter-bound fields

1.3.3 2.1 Sub-filter - Implement filter cascading

• Application scenarios: When fields with hierarchical structures are used as filters, the filter options need to display the relevant values. Prevent users from selecting distracting items. For example: province, city; primary category, secondary category.

1.3.4 Chart option configuration - Control the selectable enumerated values in filters

1.4 Analysis features #

1.4.1 Analysis features - Sorting, filtering top N values, aggregation, percentage calculation, comparative calculation, period-over-period calculation, reference lines, table calculations

• Smart Data Insight provides detailed written instructions for each analysis feature. For more information, refer to: Smart Data Insight Analysis Configuration Guide.

1.5 Other features #

1.5.1 Chart monitoring - Detect data in charts and trigger alerts

• Smart Data Insight supports the monitoring and alerting feature, which allows you to configure metric fluctuation monitoring and dimension enumerated value change monitoring.

2. Dashboards #

2.1 Dashboard configuration #

2.1.1 Dashboard themes/templates/assets - How to create more visually appealing dashboards

• For those who do not want to spend time adjusting various color schemes, fonts, or lines, try the Dashboard Theme feature in Smart Data Insight. It supports using our official themes or customizing themes according to business requirements.
• You can also use the dashboard template feature in Smart Data Insight to directly replace charts with existing templates.
• Recently, the assets feature in Smart Data Insight has recently been launched, allowing you to quickly add header images, title images, and dividers to dashboards. See Operations in the Left Control Panel Area of the Dashboard.

2.1.2 Multilingual - Set different language displays for Chinese and English environments

• In the project center, administrators can set up multilingual options for datasets, charts, and dashboards separately, enabling colleagues working in multiple languages to easily understand relevant information.

2.2 Chart interactions #

2.2.1 Chart Linkage - Linked analysis between charts

• The chart linkage feature allows a chart to be used as a filter. By clicking on a data item within the chart, the associated charts will filter and display the data corresponding to the selected item, providing a fast and intuitive experience.

2.2.2 Chart Redirect- Click a chart field to redirect to another dashboard

• This feature enables users to click a chart field and redirect to another dashboard or webpage. During redirection, it supports passing the current chart's data to the global filter of the target dashboard, or passing the chart's data to the specified location in the corresponding webpage link.

2.2.3 Navigation anchor - Quickly locate a specific chart

• The dashboard supports setting anchor points, allowing each chart to have its own anchor. Once set, the anchors are displayed on the right side of the dashboard viewing page, making it easier to quickly locate a chart when there are many charts on the dashboard. See Operations in the Right-Side Page Configuration area of the dashboard.

2.2.4 Chart dataset replacement - Quickly switch the underlying dataset of a chart

• Intelligent Data Insight supports the chart dataset replacement feature, allowing users to replace the underlying dataset of a single chart or replace the datasets used by all charts in the entire dashboard.

2.3 Dashboard applications #

2.3.1 2.2.1 Bookmarks - Save the current dashboard filter configuration as a bookmark

• Usage scenario: Users can save frequently used filter conditions as bookmarks for quick access. It also supports publicly displaying common departmental filter criteria for all users within a project.

2.3.2 Dashboard subscription - Schedule dashboard content delivery to individuals or group chats

• The subscription feature of Intelligent Data Insight allows dashboard content to be delivered as screenshots to individuals or Lark group chats. Configuration documentation: Dashboard subscription.
• Subscriptions can also be used in conjunction with bookmarks, enabling users to deliver dashboard content based on different filter conditions.

2.3.3 Dashboard/chart embedding - Embed dashboards or charts into documents or data portals

• This can be achieved through the dashboard and chart embedding feature.

3. Datasets #

3.1 Public data - Add Chinese calendar, public opinion data, weather data, and more to datasets #

• Intelligent Data Insight provides public data such as the Chinese calendar, public opinion data, weather data, and more, which can be combined with datasets to facilitate analysis.

3.2 Multi-day synchronization - Synchronize partitions for the current day as well as for the previous N days #

• Intelligent Data Insight datasets support the multi-day synchronization feature. Multi-day synchronization means that when synchronizing a specific partition, the partitions for the previous N days will also be synchronized at the same time. See Synchronization Settings.

3.3 Mirrored datasets - Enable cross-project use of datasets #

• Dataset producers can use the mirrored dataset feature to, through simple configuration, allow multiple projects to use the same dataset (same data source, data model, dimensions, and metrics) for creating charts and dashboards, saving storage resources and reducing operational costs.
• Common use cases: The same dataset needs to be reused across multiple business lines, but unified data management and operations and maintenance standards are required.
• Datasets support cross-project use, but dataset management permissions are required for configuration. It is possible to make the current project's dataset available to multiple projects without occupying additional storage space. The datasets in these projects will remain consistent with the original dataset. For configuration, refer to the help documentation: Mirror Dataset.

3.4 Associated datasets - Enable direct join queries across multiple datasets #

• Intelligent Data Insight supports the Joined datasets, allowing multiple datasets within the same cluster to be associated.

4. SQL queries #

4.1 Save chart - Save SQL query results to visual queries #

• The results of SQL queries can be saved to visualizations, allowing data analysis through visual queries.

4.2 Upload temporary table - Upload a local file as an online Hive table #

• SQL queries support uploading local CSV files as temporary Hive tables for use in queries.

5. Attribution decisions #

5.1 Anomaly analysis #

• Anomaly analysis is used to detect abnormal fluctuations in core metrics. Based on user configuration, the system automatically explores metrics daily or weekly, uses intelligent algorithms or configured rules to detect abnormal fluctuations in metrics, sends alert notifications, and allows further drilling down into dimensions to investigate anomalies.

Applicable scenarios:
Core business metrics fluctuate every day. How to determine whether yesterday's fluctuation was abnormal?
If an anomaly occurs in a metric, an alert should be triggered. However, simply relying on a week-over-week threshold does not take into account the metric's own trend and periodicity. How to make a reasonable determination?
The overall dashboard metrics may appear stable, but are there hidden anomalies beneath the surface (by drilling down into various dimensions)?

For information on using the anomaly analysis report, see Attribution Report - .

5.2 Dimensional attribution #

• Dimensional attribution is used to break down metric changes across different dimensions and provide quantitative explanations of their contributions. Based on user configuration, metrics are automatically explored daily or weekly, and scientific attribution algorithms are used to analyze the factors influencing the overall dashboard metrics.

Applicable scenarios:
For example, when performing DAU change attribution, which business sources, scenarios, or cities caused the DAU change, and how much did each factor contribute to the DAU change? Business source, scenario, and city are all different dimensions.

For information on using the dimensional attribution report, see Attribution Report - Dimension Attribution.

5.3 Metric attribution #

• Metric attribution is used to explore the impact of multiple related metrics on changes in core metrics, including the contribution of multiplicative factors and the contribution of each factor in non-multiplicative formula relationships. According to user configuration, the system will automatically analyze the impact of relevant metrics on core metrics on a daily or weekly basis.

Applicable scenarios:
With sales revenue as the core metric, how much do the user coupon redemption status, merchant livestreaming status, user viewing status, and product click status each contribute to changes in sales revenue?

For information on how to use the metric attribution report, see Attribution Report - Metric Attribution.

5.4 Metric Analysis Tree #

• The three analysis features introduced earlier are independent. However, actual business scenarios can be more complex, and a single analysis scenario may not be sufficient. For example, it may be necessary to attribute multiple dimensional metrics simultaneously or to link multiple attribution capabilities. In such cases, the metric analysis tree can be used.
• It breaks down our attribution capabilities into individual nodes, allowing users to organize and link different types of nodes based on their actual business scenarios. For example, it is possible to first perform anomaly attribution on a metric. When a metric anomaly occurs, metric attribution is performed on that metric, followed by dimensional attribution on these metrics.

Help documentation: see Metric Analysis Tree.