Among the input or filtering tools a dashboard user should have are selection lists, calendars, and range sliders. Simply by clicking or dragging, user should be able to narrow down the data view of interest to them. Being dual-purpose means that the input tool shows you the available choices before the user even begins to modify the view. In other words if the data range of available data in the dashboard is only 2 years, then the date calendar will only show that by coloring the possible choices. Another example would be in selection list. A selection list may only show two items (out of many) that are filled squares. This indicates that the result set includes only data from these two choices. The remaining choices are listed under 'Others'.
A selection list should rovide either single-selection or multiple-selection capability, and can be configured to submit selections immediately when you make the selection or on-demand when you press the 'Apply' button . To filter data using a selection list, select the desired item(s) in the list. If an 'Apply' button is visible in the title bar, press the 'Apply' button to submit your selections. If you want to make only a single selection (all selections being mutually exclusive), hold down the Alt key on the keyboard and make your selection. This changes the Selection List to "single-selection" mode. Alt -click again to return to multiple-selection mode.
The most common way to represent a key performance indicator (KPI) is by using a gauge component for graphical representation or a text component for digital representation. A gauge could be large numbers or arrows color-coded by condition such as green for good or up, andr red for bad or down. Other types are speedometers, donuts, thermometers, and bullet charts.
A gauge component allows you to display a single aggregated value (for example, a grand total) in a visually prominent manner. In addition to displaying static text, a text component allows you to display a single aggregated value (for example, a grand total).
The easiest way to use a Gauge or Text component to display a KPI is to bind the Gauge or Text to an existing data field as described in the previous sections. However, you can also compute a KPI using script. There are several ways to do this: create a calculated field, and bind the calculated field to the gauge or text in the usual way. Or enter an expression for the 'Text' property on the general tab of the 'Text Properties' panel to aggregate the desired data. Or enter an expression on the Script tab of the 'Gauge Properties' or 'Text Properties' panel to compute a desired value.
With the click of a mouse, users should be able to sort data tables by column from largest to smallest or vice versa. Chart axes should be similarly sortable. To sort chart groups based on a measure, users should hover the mouse over the desired measure axis. This displays the 'Sort' button next to the axis. Press the button once to sort the groups in ascending order by measure value, press a second time to sort in descending order by measure value, and press a third time to restore the original order. Developers should also be able to set a dashboard to display and sort values for Top-N rankings.
A target line is a horizontal or vertical line drawn on the chart that denotes an ideal value (goal or threshold) or representative value (average, minimum, etc.). A target band is a horizontal or vertical band drawn on the chart that denotes either an ideal range (e.g., goal zone) or representative range (e.g., span of maximum to minimum). A statistical measure is a line or region drawn on the chart to represent one or more statistical quantities derived from the data (confidence intervals, percentiles, etc.). A trend line is curve that is fit to the data based on a selected model (linear, quadratic, etc.).
When you hover the pointer over a data group on a chart, the tooltip provides useful information about the dimensions and measures represented by the group. You can also choose to replace the tooltip with a data tip. A data tip displays information about a data group using another Dashboard component, such as Chart, Text, or Gauge (or a collection of such components). The data tip is dynamically filtered to show only data corresponding to the particular chart group on which you hover. You can create a custom tooltip by using HTML markup to style the tooltip text. If you expect end-users to make changes to the chart using the Ad Hoc Editing feature, you should use field names rather than indices. Indices may not work correctly after a user makes ad hoc changes to the chart.
End users should be able to add annotations to a dashboard, to an individual dashboard component, and even to an individual data point in order to share information among viewers or even just a reminder to themselves. The annotation could explain why there is an outlier, if there was an outage or a holiday, or it could impart some interesting finding or call-to-action. Adding an annotation should be as easy as right-clicking in an empty region of the dashboard, and selecting 'Add Annotation' from the context menu.
A bookmark allows you save your current Dashboard settings (filter selections, menu choices, annotations, etc.) so that you can return to these settings at a later time. This makes it possble for novice users to customize their views of a pre-built dashboards. Bookmarks should also be shareable so an interesting data view can be shared without the recipient having to follow the same customization and filtering steps that the bookmark maker did. There should be a 'Read-only' option if you do not want other users to modify this bookmark. And there should be a default bookmark for a dashboard which specifies the settings (filter selections, menu choices, annotations) that will be in effect each time you re-open the dashboard in the portal, i.e., the "starting state" of the dashboard.
With a simple click-to-share button users should be able to share a dashboard to Facebook, Google Hangouts/Chat, LinkedIn, Twitter, Slack, or just email or copy link. Sharing on Facebook should not require the application to be accessible from the internet.
Examples of what-if scenarios are "What if I sell only 97% of the product I wanted to sell this month? How will that affect my bottom line for the quarter?"; "What if our trucks get snowed in this winter? What will that do to our supply chain?"; "What if my sales people outperform their quotas? How do I have to modify my budget to pay their commission?"
Machine Learning Integration
The software productionalizes machine learning. It provides visual tools that directly tap into machine learning business applications. Business users can not only visually analyze machine learning results but also train machine learning models with sliced-and-diced data. When data grows beyond an experimental machine learning setup, the native Apache Spark based platform can easily scale on demand. It can utilize an external Spark Cluster or build its own with minimal Spark expertise. Learn more about the benefits of integrating ML and dashboards.
Ad Hoc Wizard
An Ad Hoc Wizard is a simple guided procedure that helps a user to create a basic report featuring a chart, table, or crosstab, depending on the Wizard that the user selects. The user can then proceed to make further modifications to the report using the Ad Hoc tools.
Wide Array of Data Access Options
The dashboard platform should be able to connect to any relational database as Microsoft Access, MySQL, or Oracle. It should be able to access many cloud-based sources such as Google Analytics, Ads, Facebook, salesforce.com, and AWS. It should permit the import of user or externally generated Microsoft Excel spreadsheets and comma-delimited text files. For a longer list of supported data sources, see https://www.inetsoft.com/solutions/all_data_sources/
Global Styling
Developers and administrators should be able to make format choices like colors and fonts uniform across dashboards to facilitate a corporate theme or especially when embedding dashboards into a cloud-based vertical product for a consistent brand image seen by clients. Developers should be able to deploy CSS files to set styles for all dashboard components. In the same way that global style options allow you to change the appearance of your dashboard without any effort, predefined dashboard themes provide an easy way to customize the look and feel of the dashboard.
Universal Device Support
A dashboard should display properly on all device types from desktops and laptops to tablets and mobile phones. No client app should be necessary. Only a browser. This maximizes reach and minimizes support overhead. HTML5 is the primary solution for how this is accomplished. The server application renders the dashboard in HTML5 which supports all the rich interactivity that would be expected.
At the same time the platform should provide the option to customize how a dashboard is displayed on an electronic device and on the printed pages to optimize the user experience or even load time. A Dashboard "layout" identifies a class of devices (e.g., phone, tablet) based on screen dimensions. The Dashboard layout allows you to tailor the display of the Dashboard to this class of devices.
Data Mashup
Data Mashup, at its basic definition, means combining disparate data sources that were not previously architected to be used together. The key benefit is to get a complete view of organizational performance or to be able to include all relevant data in an analysis.
In a data warehouse environment with rigid ETL process, data mashup yields other benefits, particularly in allowing prototyping of data manipulations before committing to new, official ETL definitions and transformations.
BI professionals should be able to easily mashup disparate data into analytic data blocks and also enable controlled self-service data mashup for users via data models. Data models map prepared data to business terms and completely shield end users from underlying technical details. Data blocks' built-in visual transformation and data cleansing functions make data preparation effortless with minimal technical skills.
Models' built-in governance enforces data consistency and safeguards data processing engine integrity. As part of data governance, the data security model can secure data down to the cell level where users will only access their personalized data.
 |
Read more about InetSoft's data mashup technology to learn how it makes BI asset delivery and maintenance more efficient than traditional BI solutions. |
Data Caching
For cases where data transfer is slow or costly, or operational data source impact needs to be minimized, or when many on-the-fly data transformations and calculations can slow the rendering of a dashboard, there should be an automated data caching option. The preferred approach is the materialized view. Materialized view is a caching strategy that pre-aggregates and stores the data required by a dashboard. When a dashboard has a materialized view, the dashboard will attempt to query the cached materialized view rather than the database.
This can significantly reduce the run time for certain queries, and greatly improve performance for the end-user. Dashboards that possess a materialized view will rely on this materialized view to supply their data. If no materialized view is found, dashboards will either query the database directly or automatically trigger the creation of a materialized view. To improve performance, you can distribute a materialized view across multiple machines.
You can schedule a materialized view to be regenerated on a predefined schedule by assigning a "cycle" to the view. (See Creating a Materialized View for more information.) This allows the materialized view to be updated with new data from the database. However, for a large data set, repeatedly regenerating the materialized view in its entirety may be inefficient. To improve performance, you can elect to update the materialized view incrementally based on a set of specified conditions. For example, you can selectively update the materialized view with data which has posted subsequent to the previous materialization. Likewise, you can delete selected records from the materialized view if they meet a specified condition, such as an expiration date. You can specify an incremental update condition both for mergeable queries (e.g., relational database queries) and for non-mergeable queries (e.g., non-relational database queries, un-parsable queries, etc.).
Clustering
To improve dashboard performance, you should be able distribute materialized views or data caches across multiple machines. This distributed configuration is called a Spark cluster. In a Spark cluster, the materialized dataset is saved on separate physical machines (nodes) in the grid. This allows the dashboard server to pull small blocks of data from each of the nodes, thereby distributing the processing load and minimizing the workload for any particular machine. This increases total throughput and improves Dashboard responsiveness.
A Spark cluster is completely transparent to the end-user. Users can access their dashboards in the same manner regardless of whether a Spark cluster has been configured or not. However, for large data sets dashboard performance is often improved with a Spark cluster in place. Distributed materialized views utilize the Hadoop distributed file system (HDFS) together with a Spark cluster. The Spark cluster provides ready scalability by adding additional machines and excellent fault tolerance.
Fine-grained Security
Security is an important part of any enterprise application. The security model provides security at both a component and functional level. The security model also provides security at a very granular object level, e.g., a user may not have access to all the data models or queries. The security model is highly adaptable and configurable, e.g., you can import your existing users/groups/roles from a database or an LDAP server without redefining them in the dashboard platform.
A security provider is a module which handles security operations for the report server. You can specify independent security providers to provide the following functions: You must specify a security provider in order to generate server audit records. Authentication: Retrieving users, roles, and groups; authenticating logins. Authorization: Storing and retrieving permissions for different components and functions. You can define a chain or authentication providers and a chain of authorization providers.
The LDAP security provider is useful when your users, groups, and roles are already defined in an LDAP environment such as Active Directory or any generic LDAP implementation. The LDAP server can only used for authentication (i.e., verifying user credentials), not for authorization (setting permissions). All user, group, and role information is then retrieved from the LDAP server and cannot be created or configured in dashboard server.
Server Logging
The dashboad server should provide the ability to log messages generated during server and scheduler execution. Administrators should have the option to use custom log configuration. Options should include specifying the 'Log File Directory' to store the log files and 'Log File Pattern' (regex) to identify the log files so that they can be displayed in admin console.
Multiple levels of detail levels are available: 'Debug' provides internal execution-tracing information for debugging purposes. This may be needed when diagnosing a problem in collaboration with Technical Support. 'Info' reports information about the state of the server, such as the security provider in use and the location of the sree.properties file. 'Warning' reports errors that do not prevent normal operation, but indicate a possible problem that should be investigated. 'Error' reports issues that prevent normal operation. It is recommended that you set the detail level to 'Warning' or 'Info' for regular use.
Docker/Kubernetes Deployment
Kubernetes is a container orchestration system that provides the following benefits: Automated (consistent, and repeatable) deployment of containers. Automatic scaling of containers. Automatic system and resource management, including storage, networking, and security. Support for DevOps for easy integration into CI/CD pipelines Support by many cloud infrastructures, including Amazon (EKS), Azure (AKS), and Google Cloud (GKE).
Multi-Tenancy
In certain deployment scenarios, particulary when dashboards are embedded into a cloud-based application serving many clients, you may need to provide different groups of users (i.e., "tenants") with access to different sets of data. This is called a multi-tenant environment. For example, if you are serving user groups from different departments or organizations, you may need to give each tenant access to data stored in a unique database or schema (different login permissions, etc.). To facilitate design of a multi-tenant environment, Style Intelligence allows you to define independent data connections for each tenant. Each tenant can access only the unique connections for which they have privileges. Additionally, you can define a common set of data that is accessible to all tenants.
You can set appropriate aliases for reports, folders, Data Worksheets, and Dashboards. You can use aliases to provide similar names to multiple tenants (e.g., a "Marketing" folder) while maintaining distinct names for administration (e.g., "Marketing-Company A", "Marketing-Company B", etc.). See the related links for details on setting aliases.
 |
Read why choosing InetSoft's cloud-flexible BI provides advantages over other BI options. |
Archiving
When a report is generated, the dashboard engine retrieves fresh data from the database regardless of whether the same dashboard has been generated previously. In an environment with many large dashboards, you may choose to pre-generate and archive these large dashboards to avoid delays associated with repeated regeneration. Users can access archived dashboards through the Portal in the same way as regular dashboards, with access controlled by assigned permissions.
There are two ways to archive a dashboard: A user can save a live dashboard into the archive by selecting the 'Save as archive' option in the 'Save As' dialog box within the Portal. A scheduled task can create an archive dashboard. A new archive dashboard will overwrite any exiting archive dashboard with the same name. The archive can be backed up an archive into a Zip file.
Auditing
The audit feature tracks report and data access, as well as asset dependencies. When auditing is enabled, all report and database access is logged in the audit database. Note that auditing requires you to configure the audit database. Passwords for external services such as databases and LDAP servers are securely encrypted using a password-derived key.
Section 508 Accessibility Compliance
The User Portal is compliant with Section 508 of the federal Rehabilitation Act of 1973. This means that most features of the User Portal are accessible via keyboard commands and corresponding voice commands.
Single-sign On Flexibility
Single Sign-On (SSO) is the process of transparently passing user credentials from one application to another so that the user is only required to present login credentials a single time. This helps to create a seamless user experience. You can implement single sign-on (SSO) by injecting a principal object into the HTTP request object as part of the security filter chain. This allows the server to automatically sign in the user without attempting to authenticate credentials.
Language Support
You can use a simple text mapping to localize most text in a dashboard, as well as data source names, data model names (entity/attribute), and query (column) names. When users log into the User Portal or administrators log into Enterprise Manager they can specify a locale (language-country combination), based on which the server dynamically modifies the interface (Portal tabs, Enterprise Manager tabs, repository tree) as well as report and Dashboard elements (text elements, chart axis titles, table header columns, etc.).
Embeddability
A complete dashboard and reporting solution needs to provide the right mix of features and ease of use, but also must fit into the host application as seamlessly as possible. This means that the dashboards and reports must be transparent to the end-user and must look and feel like a normal part of the host application - whether it is used by either internal users or external customers.The dashboard software should be 100% open standards based and rely on these open standards such as AngularJS, REST, and Auth0 to ensure tight integration. Key features include: Single Sign On, Custom Theming, Wide Variety of Data Sources, and a Single Web UI for all Devices. If you are an application developer, be sure to read our top embedded BI features article.