Data Explorer
Visually inspect and compare dataset variables using interactive charts and statistics.
Overview
The Data Explorer lets you visually inspect variables from your datasets, uploads, or external data sources. It provides four views — individual variable trends, multi-variable comparisons, scatter plots, and descriptive statistics — all filterable by modeling combination.
Source Selection
Before exploring, select a data source (dataset or upload). This populates the available variables and modeling combinations. You can then pick which variables to inspect and optionally narrow down to specific Country/Region/Grouping combinations.
Tabs
Each Variable
Displays one sub-tab per selected variable with two visualizations:
- Explore by Date — a line plot of the variable's values over time.
- Country-Region-Grouping Breakdown — a sunburst chart showing the variable's distribution across combinations (unaffected by the combination filter).
Variables Comparison
Shows all selected variables on a single line plot over time. Hover over a date to see a breakdown in an accompanying bar and pie chart.
If two or more variables are selected, a correlation matrix is also displayed below the line plot.
Scatter
Plots variables against each other. If more than three variables are selected, a secondary selector lets you pick up to three for the plot. The chart adapts automatically:
- 1 variable — distribution plot
- 2 variables — standard scatter plot
- 3 variables — 3D scatter plot
If the selected variables have no overlapping dates, a message is shown instead.
Statistics
A sortable table with one row per variable showing: Max, Min, Sum, Mean, Median, Standard Deviation, and Range.
Aggregation
On the Variables Comparison and Scatter tabs, an aggregation selector controls how values are combined across modeling combinations. Options: sum, mean, min, max.
Tips
- Start with the Each Variable tab to spot anomalies before comparing variables.
- Use the Statistics tab for a quick sanity check on data ranges and distributions.
- Narrow down to a single modeling combination when comparing variables to avoid aggregation effects masking patterns.
See Also
- Activities — track modeling activity across your project
- Filtering — learn about combination filtering