Ask the question where you cleaned the data
The analysis stage is in the same platform as the preparation stage on purpose. When a chart looks wrong, the transform history that produced it is one click away, rather than in a different tool owned by a different team.
Explore and visualise
Charts
Bar, line, area, scatter, pie and distribution charts over any dataset in the catalogue, with the chart type suggested from the column types you picked.
Pivots
Group, aggregate and cross-tabulate without exporting to a spreadsheet and losing the lineage.
Summarise
A written summary of what the data shows, generated against the actual profile rather than against the column names.
Dashboards
Assemble the charts that answer a recurring question into one surface.
Simulate and decide
Exploration tells you what happened. Simulation asks what would happen. You define a scenario — change these assumptions, apply this growth rate, remove this segment — and compare the outcome against the baseline and against other scenarios side by side.
Because it runs against the governed dataset rather than a spreadsheet copy of it, a scenario stays valid when the underlying data is refreshed, and the assumptions behind it are recorded with it.
The assistant, in the analysis stage
It has the profile
Ari answers about your columns, distributions and row counts, not about data in the abstract.
It shows its work
An answer comes with the aggregation behind it, so the number can be checked rather than trusted.
It exports
Any answer can be copied out or exported as a document, with the tables intact.
Analytics questions
Does this replace my BI tool?
For exploratory work and for the analysis that happens while data is being prepared, yes. For a company-wide reporting layer, DataLens more often feeds one — publishing a governed data product that the BI tool reads.
Can I share a chart with someone who does not have an account?
Yes, through an access-controlled share link. Sharing is a governed action, so what was shared and with whom is recorded.
What is the difference between Explore and Simulate?
Explore interrogates the data as it is. Simulate changes assumptions and compares outcomes. They are separate surfaces because they answer different kinds of question.
See it on your own data
DataLens is in private beta. Bring a file, a database or an API and work through the whole lifecycle in one sitting.
Request beta access