Okay, now let's get a reality check with the product and see which features it is either lacking or needs significant enhancement, so that it can be considered a viable enterprise BI tool replacement.
In the next few pages, I will group these features in categories and also indicate which of these features are already in the product roadmap.
The following are the features that I expect will be added to the product for integrations:
- Integrations with SPICE (in roadmap): SPICE is what makes QuickSight really fast and is designed to run interactive queries on large datasets in memory. SPICE is currently only being used by QuickSight visualizations and the product management team has mentioned that external applications such as Tableau, DOMO, and Qlik will be able to integrate with this layer, making the QuickSight value proposition really higher. It is also suggested that SPICE will have a SQL-like access provided for additional third-party applications via APIs and/or an ODBC connector.
- Embedding visuals to other websites (in roadmap): If you, as a user, like the reports generated by QuickSight, you might need the ability to embed it on a custom website. This feature should support passing variables as parameters to the dashboard to make this integration easier to use.
The following are the features that I expect will be added to the product for the category of visualizations and analysis:
- Clone existing analysis: Often, you may wish to preserve an existing analysis and create, and then modify, a new one based on it.
- Advanced charting: The charting options provided by QuickSight do support the commonly used scenarios, but there are several other types that enterprise BI applications have that are missing. The following are a few:
- Geospatial heatmaps (in roadmap)
- Gantt chart
- Histogram chart
- Filled maps (colors based on a metric on a region/country)
- Box-and-whisker plot
- Regression analysis
- Source code management: To promote code from one environment, such as development to production, we need the ability to check-in and check-out the analysis and dashboard definitions to a Git or SVN repository.
- Site branding (in roadmap): QuickSight currently does not allow users to customize the look and feel to suite their custom corporate needs like a logo and custom CSS.
- Usability: QuickSight needs improvement in managing filters, legends, and its ability to add text to analysis.
This completes the feature gaps in visualizations and analysis; next, we will review the gaps in sharing and collaboration.
In this section we will review the features lacking in the sharing and collaboration category:
- Mobile support for android devices (in roadmap): Currently mobile support is limited to Apple devices and I expect the android native app to be released soon
- Download data and charts (in roadmap): From the analysis and dashboard pages, there is no option to download the data and/or visuals to Excel or PDF for offline analysis and sharing
- Usability: With dashboard feature, the product needs improvement in managing filters, adding text, and images, to make it more effective
The following are features that I expect will be added to the product for the category of data management:
- Integration with DynamoDB (in roadmap): DynamoDB is a popular NoSQL database service provided by AWS and currently there is no integration with QuickSight.
- Hierarchy support: QuickSight charts do support drill down so that we can go from a higher level of metrics like state level to a child level like city. This hierarchy management is not easy to use and is useful in several types of dashboards/metrics reporting.
- Organizing the catalog: Currently, the list of existing data sources is shown under the Manage data section. This list is a simple tile view with three data sources shown per line ordered by last used on top. This should be better organized by category and importance by the end user. Additionally, each tile should show a quick stat like total number of records and size of the source.
The following screenshot shows the current catalog which simply orders all data sets in descending order of creation:
Figure 8.4: Data sets not organized
Let's review the data preparation features that QuickSight is lacking:
- Operations pipeline: When you revisit a data set which had a series or data preparation steps, it is not clear as to what changes were made and in what sequence. Other BI tools typically show the list of operations either as steps or visually as a pipeline.
- Lineage: QuickSight supports joins and functions which alter the original data. For compliance and audits, it is necessary to have a lineage report to show where the data came from along with all the operations done on it at every field level. This lineage feature does not exist today.
In this section, we will review the features lacking related to security:
- Governance (in roadmap): QuickSight has plans to add governance to data sets so that an administrator can manage who gets access to which data sets
- Fine grain access control (in roadmap): QuickSight does not have the ability to restrict certain portions of data to specific individuals; for example, for compliance needs customers often want to grant access to sensitive data to only certain individuals of the organization
Here are some other general features that QuickSight lacks:
- Error logging: QuickSight server side logs are not accessible even with administration role. This would be useful in troubleshooting issues that cannot be resolved on the user interface.
- Audit logs (in roadmap): Currently there is no easy way to track who accessed what data using QuickSight.