This blog post was contributed by Data Management and Analytics Consultant Gio Schulte.
As a data management and analytics consultant, I have developed dashboards in a majority of the popular BI tools from Tableau to PowerBI, as well as their backend data structures. Earlier this year, the opportunity to develop dashboards in Looker arose when a new client needed Aptitive to help them model their data and develop Looker dashboards for their sales team (more details here).
Based on my limited experience with Looker, I knew that it makes creating quality visuals simple and that coding in LookML is unavoidable. I worried that LookML would be extremely nuanced and I would lose time troubleshooting simple tasks. I could not have been more wrong on that front. Along with this realization, below are my top takeaways from my first Looker project.
Takeaway 1: LookML is easy to learn and ensures consistent metrics across reports.
Given the vast amount of documentation provided by Looker and the straightforward format of LookML code, I quickly caught on. This learning curve may be slightly different for report developers who have minimal experience with SQL. LookML adds transparency into what happens with data presented in visuals by directly showing how the code translates into the SQL queries that run against the source data. This makes it much easier to trust the results of dashboards and QA as you develop.
More importantly, LookML allows users to ensure that their metric definitions are consistent across dashboards and reports. Establishing this single source of truth is key for the success of any reporting efforts. Within the semantic layer of the reporting tool, users can create SQL queries or harness LookML functions to develop custom measures and include descriptions to define them. Transforming the source data into predefined measures in the back end of the reporting tool ensures that report developers access the same metrics for every dashboard business users will see. This is a clear contrast from tools like PowerBI and Tableau where the custom measures are created in each workbook and can vary. Furthermore, by using roles, administrators can limit who has access to change this code.
Takeaway 2: Creating dashboards and visuals is super intuitive for about 95% of use cases.
After setting up your data connections and LookML, developing a visual (“Look”) in Looker only requires a simple point and click process. Once you select the filters, measures, and dimensions to include in a visual, you can click through the visualization options to determine the best possible way to present the data. From there, you can easily adjust colors and stylistic options in settings using drop down menus. Compared to other BI tools, these visuals are fairly comparable and standard across the board. That being said, Looker greatly stands out when it comes to table visualizations. It allows for conditional formatting similar to that in Excel and a wide range of visual options in comparison to other BI tools. This makes Looker a great selection for companies that often require tables to meet reporting requirements.
Although detailed documentation and the simple interface meet most reporting needs, there are limitations when it comes to easy customization in visuals. This includes the inability to set drill ins by a visual rather than a field. In Looker, any demographic used across reports has to drill into the same fields (unlike those set per visual in a Tableau Tool Tip, for example). Additionally, you cannot format visuals based on customized metrics (ex: color bands, conditional formatting for Field A based on the value of Field B, etc.). The caveat here is that you can unlock many customized visuals by writing custom code, a skill not always handy for report developers.
Takeaway 3: Looker is extremely collaborative, something not often seen in BI tools.
With most BI tools, developers are forced to work independently since two people cannot easily contribute to a single workbook at the same time. Looker’s web based format seems to have been built with collaborative development in mind, making this tool stand out when it comes to teamwork. Business users can easily contribute as well since the web based tool makes sharing dashboards and embedding them within websites easy. While this may seem minor to some, it significantly enhances productivity and yields a better end product.
The following features ensure that your team can iterate on each other’s work, edit the same dashboards and develop LookML without accidentally overwriting work or creating multiple versions of the same report.
- Version control and deployment processes are built into the “Development” window where users can modify and add LookML code.
- Ability to duplicate Looks developed by others and iterate on them. These Looks can then be added to dashboards.
- Shared folders where Looks and Dashboards used by multiple people can be stored and reused (if needed).
- Ability to “Explore” a Look created by someone else to investigate underlying data.
- Ability to edit a dashboard at the same time others can make changes.
- Sharing dashboards using a link and the ease of embedding dashboards allows for seamless collaboration with business users as well.
With a properly modeled data source, Looker impressed in terms of its performance and ability to provide highly drillable dashboards. This enabled us to dramatically reduce the number of reports needed to address the wide range of detail that business users within a department required. While the visuals were not as flashy as other BI tools, Looker’s highly customizable table visualizations, row level security, and drill in options were a perfect fit for their use cases.
Aptitive specializes in advising companies on how to gain the most business value possible from their analytics tools. We assist organizations with everything from selecting which tool best suits their needs to developing dashboards for various departments or structuring data to enable quick reporting results. Contact Aptitive if you need help determining if Looker is the tool you need or want guidance on how to get started. Aptitive can help get you on track to better analytics with our Data Visualization Starter Pack.
Originally published at https://aptitive.com on November 4, 2020.