> ## Documentation Index
> Fetch the complete documentation index at: https://help.gowindmill.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Explore
> Build your own analytics view by picking any stats Windmill collects, filtering by employee, and charting the result however you want. Export the underlying data to CSV when you need it.
Explore is Windmill's ad-hoc analytics builder. If the [AI Adoption](/features/analytics/ai-adoption) and [Product Usage](/features/analytics/product-usage) reports are pre-built dashboards, Explore is the blank canvas. Pick the stats you care about, scope to the right people, and visualize the result.
Use Explore when you need to:
* Answer a specific question that the pre-built reports don't cover ("How many 1:1 notes did the Sales managers edit last quarter?")
* Compare a few stats side by side for one person or team
* Export raw stat data to a CSV for further analysis
## Building a query
A query in Explore has three parts: **stats**, an **employee filter**, and a **date range**. Once you've picked at least one stat and the filter resolves to at least one person, Explore renders a chart and a data table.
### Picking stats
Click **Select Stats** (on the empty state) or **+ Add stat** to open the stat picker. The picker is organized by category — activity, code, AI tools, 1:1s, feedback, meetings, Pulse, performance reviews, and more. You can pick as many stats as you want.
Selected stats appear as removable chips at the top of the page. Click the × on a chip to remove it, or use **Clear All** to reset.
For a full catalog of every stat Windmill tracks and how each one is defined, see the [stats reference](/data/stats).
### Filtering employees
Open the **Employee filter** to scope the query. You can stack any combination of conditions — for example, "Department is Engineering AND Manager is Jane AND Employment Type is full-time." Leave the filter empty to include everyone in your visible scope.
If your filter resolves to a single person, the data table changes shape (see [Tables](#tables) below).
### Date range
Use the **Date range** picker to set the window. Quick presets: **last 7 days**, **last 4 weeks**, **last 12 weeks** (default), **last 6 months**. Custom ranges are supported.
## Chart and table view
Once you have stats and employees selected, Explore shows two things: a chart on top and a data table below.
### View modes
Use the **Trends / Employees / Employee Trends** toggle to change what the chart's x-axis represents:
* **Trends** — x-axis is time. Each series is a stat (or aggregated across stats if you've picked multiple). Best for "is this metric going up over time?"
* **Employees** — x-axis is each employee in scope. Best for ranking — "who edited the most 1:1 notes last quarter?"
* **Employee Trends** — x-axis is time, but each employee gets their own line. Best for comparing a small group of people over time.
When you switch to **Employees**, the chart type defaults to **Bar**. Other modes default to **Line**.
### Chart types
Choose between **Line**, **Bar**, and **Stacked Bar**. Stacked Bar is most useful in Trends mode when you have multiple stats and want to see their combined total per period.
### Granularity
In time-based views (Trends and Employee Trends), use the **Daily / Weekly / Monthly** toggle to control how dates are bucketed. Choose the granularity that matches the cadence of the activity you're measuring — daily for high-volume signals, monthly for things that happen less often.
There's a 20-series limit on Employee and Employee Trends charts. If your filter resolves to more than 20 employees in those modes, the chart hides itself and shows a callout. Narrow your filter, or switch to **Trends** mode to aggregate across all employees instead.
## Tables
The data table below the chart adapts to your view:
* **Multiple employees selected** — Rows are employees, columns are stats. Sort by any column. Click a value to drill into that employee's trend for that stat.
* **Single employee selected (with a time-based view)** — Rows are dates, columns are stats. Sort by any column or by date.
Tables always reflect the same employee filter, date range, and stats as the chart above.
## Exporting to CSV
Click **Export CSV** in the top right to download the raw underlying data. The CSV has one row per employee × stat × date, with columns:
* `ID` — Windmill employee ID
* `Name` — Employee display name
* `Job Title`
* `Stat` — Stat label
* `Date` — Date in `YYYY-MM-DD` format
* `Value` — Numeric value for that employee, stat, and date
If an employee has no value for a given stat × date, the row is written with `0`. Use this when you need to do your own analysis in a spreadsheet or BI tool.
## Sharing a view
Your full Explore configuration — stats, filters, date range, view mode, chart type, granularity — is saved in the URL. Copy the URL to share a specific view with a teammate. They'll see the same configuration as long as they have permission to view the underlying employees.
## FAQs
Every stat Windmill tracks — activity (Windmill active days, messages to Windy), 1:1s (note edits, agenda topics, prep responses, meeting time), feedback (given, received, Windy-prompted), Pulse (responses, surveys created), performance reviews (drafts, answers, completion), code activity (commits, lines, PRs), meetings (count, time), and per-integration stats from connected AI tools (Claude, Codex, Cursor) and code tools.
Open the stat picker to see the full list grouped by category.
Employee-axis charts (**Employees** and **Employee Trends**) cap at 20 series for readability. If your filter resolves to more than 20 people, switch to **Trends** mode to see the aggregate across all of them, or narrow the filter.
Not yet. The view is saved in the URL though, so bookmarks work — and you can paste the URL into a doc or share it with a teammate.
Explore is a visual builder for browsing stats inside the dashboard. [MCP](/features/mcp) is for asking Windy natural-language questions about your data ("Summarize Engineering 1:1 frequency last quarter"). They draw from the same underlying stats — use whichever fits the moment.
Yes. Explore (and the rest of Analytics) buckets dates in UTC, so the dates in the CSV represent UTC days. Keep that in mind if you're correlating against data from another system that uses local timezones.