Overview
Explorer is the answer to "our dashboards don't have that specific view I need". Where dashboards are curated and opinionated, Explorer is a blank canvas. Pick a data source, choose dimensions and metrics, add filters, run the query, and export the result. If you prefer SQL, Explorer gives you full SQL access to the same Athena-backed data.
The right mental model: dashboards handle the 80% of questions you ask repeatedly; Explorer handles the other 20% that are one-off, experimental, or specific to a particular investigation. Every analyst touches Explorer eventually.
Navigate to Explorer
Open Insights > Explorer from the left sidebar. Explorer has a wider layout than the dashboards because query building needs panel space for the dimension picker, metric picker, filter list, and results table.
If this is your first time, start by picking a saved query from the recent list (if any) to see how queries are structured. Building from a working query and modifying it is faster than starting from scratch.
Select a data source
Pick the Echo source or dataset you want to explore. The schema panel on the right updates to show every dimension and metric available in that source. Dimensions are the grouping columns (source, medium, country, page path, event name); metrics are the numeric aggregations (sessions, users, events, revenue).
If the source you need isn't listed, it probably isn't configured in the Insights data catalog yet. Submit a dataset request and the team can set up a new source (this is common for first-time deployments or when a client onboards new data).
Build your query
Select your dimensions and metrics by clicking them in the schema panel. Set a date range. Add filters if you want to exclude or focus on specific dimension values. Explorer builds the SQL for you in the background and shows the preview query so you can sanity-check the logic.
For advanced scenarios (window functions, joins, CTEs, multi-step logic), flip to the SQL editor and write directly. Explorer's SQL editor has schema awareness and basic autocomplete. Queries run on the same Athena instance as the dashboards, so performance characteristics are identical.
Keep date ranges tight during query development. A query across a year of data can take significant time and cost; drop to a week or a day while iterating, then expand the window once the query shape is right.
Run and export
Click Run to execute. Results appear in a sortable table below the query builder. Click any column header to sort; use the search box to filter the result set client-side without rerunning the query.
Click Export to download the results as CSV. From there, the data is yours: drop it into Excel, Google Sheets, a pivot table, a BI tool. Explorer is a query interface; downstream analysis is wherever you're most comfortable working.
Explorer queries run against Athena reading from S3-backed data, so expect query latency in the seconds rather than milliseconds. If a specific dataset or configuration isn't available, dataset requests can be submitted for new data pipelines. Results are fully downloadable as CSV for offline analysis, and Explorer is the right tool whenever the pre-built dashboards don't have the exact view you need.
Troubleshooting
Query errors out with "table not found"
The data source probably isn't loaded into Athena yet, or the table name in your custom SQL is misspelled. If you're using the visual builder, switch sources and try again. If you're writing SQL, list available tables via SHOW TABLES to confirm the exact name and database.
Query takes forever or fails with "query timeout"
Most slow queries scan too much data. Narrow the date range, add a filter on a partitioned column (usually date or event date), and run again. Athena cost and time scale with data scanned, so a tight date range is usually a 10x speedup. If the query is legitimately complex and needs all that data, consider materializing the result as a new Insights dataset instead of rerunning ad-hoc.
CSV export is missing rows that were visible in the table
Explorer paginates the in-page result table for performance but exports the full result set. If you filtered client-side before exporting, the CSV reflects the unfiltered query output. Rerun the query with the filter baked into the query itself to get a filtered CSV.