Have you ever tried to build a segment, an AI model, or a set of match rules and felt like your data was in a ‘black box’? You know the data is there, but you can’t quite tell if there are any missing values or other silent data gaps, which can be costly in downstream usage. Considering the rate at which the volume and complexity of data increase across your organization, the lack of visibility can be debilitating.
Unlocking that visibility in your data will eliminate guesswork, accelerate critical workflows, and promote confident decision-making. However, in order to activate this data, you’re often required to switch between multiple tools or run complex queries to acquire basic insights.
With that said, I’m thrilled to introduce Data Lens, an interactive data exploration tool designed to provide fundamental visibility in your data. Data Lens is embedded directly into applications on the Data 36O platform like Segmentation and Einstein Studio, empowering you to seamlessly explore and build trust in your data. This means that you no longer need to switch between multiple tools or even write any code to instantly unlock insights in your data.
I’ll start by highlighting the core vision behind Data Lens and the suite of features designed to transform your data experience. After the product overview, discover exactly where you can find Data Lens in action today. Lastly, preview the prospective use cases and upcoming integrations currently being explored for this product.
Interactive Data Profiling with Data Lens
Data Lens is built to transform the way you interact with your datasets. Our core objective is to help you investigate data in an intuitive, self-service manner, accelerating your time to value and ensuring the data you rely on is fit for purpose. Data Lens delivers these capabilities by providing sanity checks and granular insights through the Profile Pane and Data Grid.
How It Works
Data Lens automatically loads sampled data from your source to provide a representative preview of your dataset. This design ensures a responsive user experience while still giving you the insights needed to trust your data. If you want to see more or fewer results, you can easily adjust the sample size and method (e.g., random or top N rows) in some applications to adapt your data representation.
The User Interface
As highlighted above, Data Lens is built around two primary components: (1) the Profile Pane for high-level insights, and (2) the Data Grid for granular control.
1. Profile Pane: Field-Level Summaries
The Profile Pane instantly surfaces crucial diagnostics through highly interactive profiling, allowing you to quickly spot issues that might otherwise remain hidden.
- Distributions and Binning: Instead of displaying thousands or possibly millions of individual values, Data Lens automatically groups distinct data into bins. This process creates histograms that immediately reveal the shape of the data, revealing patterns, gaps, and outliers. You are then empowered to identify deeper insights such as seasonal trends, turning millions of rows into a clear and actionable summary.
- Search: Use the search bar to quickly locate and jump to specific fields within a large dataset.
- Brushing: Verify complex relationships and understand how specific fields affect distributions in another. By selecting a value in the Profile Pane, you instantly highlight related data across all fields and filter the Data Grid to show only those specific records, enabling detailed cross-column inspection.
2. Data Grid: Row-Level Granularity
The Data Grid complements the Profile Pane’s high-level summary with a granular display of individual records. This component is essential for making metadata actionable at scale, as you can spot-check individual rows to further investigate observations and sanity-check your data without writing any SQL queries.
- Sort: Sort columns alphabetically or by count to organize and examine specific distributions.
- Expanded Data Grid: Navigate to the second tab to access the expanded Data Grid, providing a spreadsheet-like table view for comprehensive inspection.
- List View Controls: Easily hide, show, and reorder fields to declutter your view and prioritize the fields you need for your current task through the third tab.
Upcoming Capabilities of Data Lens
Looking ahead, Data Lens will support quality metrics and a suite of integrated editing tools for real-time data preparation. Please note that these capabilities may not be available in all Data Lens integrations due to varying Data Lens configurations for each use case.
- Quality Metrics: Instantly identify key data quality indicators, such as the percentage of nulls, blanks, unique values, and data types. By evaluating these metrics, you can confidently take action on this data and avoid costly rework in downstream processes.
- Filtering: Right-click on a value in the Profile Pane to instantly include only that value or exclude it from your current preview, allowing for real-time data filtering.
- Inline Edits and Data Type Changes: You can directly perform basic inline edits and instantly change a column’s data type such as switching a text field to a date/time type.
- Split Values: Use the custom split feature to break out field values based on a specified delimiter such as splitting “Full Name” into first name and last name.
- Changes Pane: Data Lens keeps a log of all your modifications in the changes pane, allowing you to monitor and easily undo actions to revert to previous states.
Technical Highlights
Data Lens is built for enterprise-grade performance, ensuring a stable and reliable experience even when working with sophisticated datasets.
- Real-time Interactions: Data Lens delivers rapid distribution calculations and updates, allowing you to brush and interact with the tool instantly.
- Virtual Scrolling: Allows you to efficiently explore fields and rows within even the largest datasets by loading fields as you scroll, ensuring that Data Lens remains smooth and responsive.
Data Lens Integrations Across Data 36O
Please note that the features available within Data Lens may vary based on each application’s specific configuration.
Use Case: Data Lens in Segmentation (GA)
We integrated Data Lens into Segmentation as Segment Preview to increase your confidence in segment creation. Consider a marketing manager, Jane, building a segment to target high-value customers. Precision is crucial, as misallocated budgets and irrelevant messages compromise campaign effectiveness and erode customer trust.
- Challenge: Previously, Jane could only rely on the segment count alone. To inspect the composition of the segment output, she had to publish the segment blindly. This approach incurred unnecessary costs and created delays simply to inspect the data.
- Solution: With Segment Preview, Jane can immediately investigate the data after defining the segment rules and prevent costly rework.
- Using the Profile Pane, Jane immediately reviews the distribution of critical fields for this specific segment. She notices an unusually high percentage of nulls in the “Email Address” field and brushes by that bin to investigate if those incomplete records share any other common traits, such as being located in a specific region.
- The Data Grid displays specific rows from the segment and allows Jane to confirm the scope of the quality issue by pinpointing the exact records that are affected.
Jane conducts this iterative process to diagnose, refine, and confirm the data quality before publishing her segments. Through Data Lens, she is empowered to address issues right away, avert costly mistakes, and increase confidence in her segment results.
Please click here to read more about Segment Preview.
Use Case: Data Lens in Einstein Studio (GA)
Einstein Model Builder requires high-fidelity data to produce accurate predictions, but as you may have experienced, it is easy to overlook the silent data gaps during the initial configuration. Prior to Data Lens, it was difficult to interpret why some columns may be missing or why certain outcome variables could not be selected. Typically, these issues arise from columns missing data or outcome variables being non-binary. Data Lens eliminates this guesswork and provides instant transparency in your data prior to expending any credits for building your model. With this visibility, you can verify your column distributions and their data types more feasibly to create trustworthy and reliable models.
Use Case: Data Lens in Data Lake Objects (GA)
These capabilities are also available for Data Lake Objects (DLOs). Data Lens will provide visibility in this space to build trust and address quality concerns at the source, prior to downstream usage across Data 36O.
- Expanded Visibility: See and profile all columns in your DLO, moving far beyond previous limitations that restricted the view to only a handful of fields.
- Reduced Friction: You will no longer need to navigate to Query Explorer or write any SQL to inspect distributions and quality metrics. Data Lens surfaces these insights directly within the DLO application.
- Ensured Quality: By profiling your DLOs immediately upon ingestion, you assess data quality issues upstream.
Prospective Data Lens Use Cases
These following use cases are prospective in nature as we are currently evaluating the value and feasibility of these integrations. Nevertheless, this section provides a sense of potential incoming features. Please let us know in the comments if there are any you would like to request to be released.
Use Case: Data Lens in Data Model Objects
You likely use Data Model Objects for a variety of functions in Data 36O. Whether it’s for segmentation, calculated insights, or identity resolution, Data Model Objects are a fundamental component of Data 36O. Integrating Data Lens into this data source has a cascading benefit across all downstream applications. This preview tool will enable you to verify the source of fields and finalize mapping to the schema. It significantly reduces guesswork and prevents errors from propagating to downstream applications, enhancing your time-to-value across Data 36O.
Use Case: Data Lens in Data Transforms
Data Transforms involve complex data cleaning, aggregating, and data restructuring processes. Data Lens plays a crucial role in this space by providing a preview of your data every step of the way. You can inspect nulls, blanks, and distributions to iteratively create your transform. After you inspect the changes from each node, you may seamlessly validate and deploy your transform. By affirming your expectations of the data, you are armed to make confident decisions that prevent unnecessary credit consumption.
In addition, Data Transforms is enabling agentic capabilities where you can create, save, and run a transform in an end-to-end experience. All you have to do is enter a simple conversational prompt, requiring little to no technical knowledge. Data Lens will be a key component within this experience, providing real-time validation to build your confidence in the generated results. After identifying any misinterpretations or unintended results through Data Lens, you can save and run the transform seamlessly in one unified interface. The existing Data Transform builder alongside Data Lens and the Agent will create a comprehensive experience for business users and technical users alike.
Ultimately, these advancements enable rapid iteration and optimized data modifications while preventing costly rework. Please see the conceptual preview of the design for this integration below.
Use Case: Data Lens in Agentic Calculated Insights
Calculated Insights is a powerful tool to create metrics and KPIs on top of your existing single source of truth of customers. With Data Lens integrated directly into the CI authoring experience, you can seamlessly profile your data and preview the CI results in real time. As a result, Data Lens’ profiling speed enables you to identify and address issues in your CI definition logic quickly, increasing your confidence in the accuracy of your metrics.
Use Case: Data Lens in Identity Resolution
Identity Resolution serves a key function in Data 36O by supporting the creation of the unified customer profile. Fundamentally, this feature involves delivering clarity while unifying your data. Currently, you are unable to preview the distributions of your data prior to setting match rules. With more visibility in your data, you can identify critical setup issues such as misconfigured mappings or missing required data and then confidently set match rules to create more reliable unified profiles. Ultimately, through Data Lens, you can prevent costly match errors, verify your source data populations early, and build trust in your unified customer profile.
Use Case: Data Lens in Query Editor
Query Editor is typically used by technical users to investigate data by writing SQL queries. By integrating Data Lens and visualizing data instantly, you can iteratively debug your complex queries and rely on this tool to provide more granular insights seamlessly. By profiling the result of your queries, you are enabled to confirm your expectations of the data and are enabled to discover patterns you may not have recognized.
Get Started Today
Data Lens accelerates workflows and promotes self-service data validation, helping you quickly realize value and improve outcomes across your organization. It helps you unlock new opportunities by ensuring your insights are built on a foundation of trust, and any employee regardless of technical background can utilize this tool to derive insights from your organization’s data.
Data Lens is currently generally available in Segmentation, Einstein Studio, and in Data Lake Objects (DLO). Simply head over to the Data 36O Feature Manager and toggle on Segment Previews. Please refer to the latest Salesforce release notes for additional information and instructions for utilizing this tool in your Data 36O environment.
Get Involved: Which Data 36O workflow would you like to see Data Lens improve next? Share your feedback directly with us in the comments!
Additional Resources
Written by
