Exploring a dataset with the Data Explorer

The Data Explorer is a tool available on all published datasets. It let's you explore and visualize a dataset without having to use any complex tools.

You can filter the data and build tables, maps, and charts to:

1) Gain insight about what the dataset provides

2) Make actionable outputs that you can share with other data consumers

To use this feature, go to the asset page of the dataset you wish to explore and click on the Explore data (beta) button.

Like in the image above, the Data Explorer is divided into three main areas with the visualization in the middle of the screen.

• On the left, the "Filters" panel lets you narrow the data

• In the center, the visualization area displays the table or chart depending on the icon you select at the top: records (the default view), map, aggregated table, and three types of charts. You can switch between them at any time to look at the data from different angles, for example from a chart to the aggregated table.

• On the right, the Properties panel lets you configure it

Note also that you can toggle open or closed both the Filters and the dataset's Schema from the vertical menu on the far left. Consulting the schema can help you identify the fields you should use to filter or build a chart. To minimize or expand the Properties panel, click on the icon in the upper-right corner.

The records table is the default view when you open the Explorer tool. It shows the dataset's records in rows, with each field displayed as a column.

• To sort the table by one of the fields, click it at the top of that column, and click it again to reverse the order.

• To change how many records are shown per page, use the pagination controls below the table to the right. You can show 25, 50 or 100 records per page.

• Use filters on the left by clicking on the funnel icon, and click on the spreadsheet icon on the left to examine the schema.

To choose which columns to show and in what order, open the Properties panel by clicking on the bar to the far right:

• You can drag fields to reorder the columns, or click its menu icon and select Move up or Move down.

• To hide a column, click its menu icon and select Hide, or drag it to "Hidden in table" at the bottom of the list. To hide every column at once, click Hide all.

• To show a hidden column again, go to the bottom under "Hidden in table" and select Show, or drag it back to "Shown in table". To show every column, click Show all.

If the dataset has at least one geographical field, you can display its records on a map. Otherwise, the map icon is unavailable.

The map uses the dataset's geographical field automatically. If the dataset has more than one, a field containing shapes is preferred. To change this, open the Properties panel by clicking on the bar to the far right, and select another available field under "Geographical data field".

Under "Geometries", select which types of geometry to display: Points, Lines or Shapes. All three are shown by default.

If the field only contains one type, it's detected automatically and the choice isn't offered.

To give each category its own color, toggle "Group by category" within the Properties view:

1. Under "Categories field", select a field. Its values become the categories.

2. Under "Category list", select the categories you want to color. By default, all values are displayed.

Otherwise, by default only the 20 most frequent categories are listed and colored. Every one of those records is displayed on the map, and records outside those categories are grouped as "Other categories" in the legend.

By default, the same categories apply to points, lines and shapes. To set different categories for each geometry type, turn off "Same categories for all geometries".

Note that you can click a point, line or shape on the map to see its corresponding record.

An aggregated table is a table grouped by a field you choose, with one or more calculated columns. An aggregated table therefore starts empty: there's no default grouping or calculation, so you configure both before anything appears.

1. Under "Row grouping," select a field. Its distinct values become the table's rows.

2. Under "Values," select a function — Average, Count distinct values, Count records, Maximum, Median, Minimum, or Sum. Then, for any function except Count records, the field to apply it to.

You can add up to 10 calculations, each becoming its own column.

To help you get started, note that the fastest way to see results is to group by any field and add a "Count" calculation. Count needs no target field, so it populates the table right away.

To build a chart, filter for the data you wish you analyze, choose a style of chart and which specific type of that chart you want to make, then configure the fields and calculation you want to use.

Use the Filters panel to refine the data shown in the chart.

As you add or change filters, the visualization updates to reflect the selected subset of the dataset. This makes it easier to focus on the values, categories, or time periods you want to analyze.

Explore data supports three chart families, and each chart family includes one or more sub-types:

• Column charts can be configured as both columns and bars

• Line charts can be either simple lines or area charts

• Pie charts can be configured as pie or doughnut charts

See "Multiple series and stacking" below for how to create stacked columns, bars, and area charts.

When configuring a chart's properties, the available options depend on the chart type and sub-type you select. As you update the chart properties, the preview updates automatically.

• the X-axis and Y-axis

• the Field, Sort by, and Order values

• for one or more series, their Function and Field

• for a pie or doughnut chart, the Function and Value

The available functions are:

• Average

• Count distinct values

• Count records

• Maximum

• Median

• Minimum

• Sum

The most common source of confusion is the difference between Count records and Count distinct values:

• Count records counts every row in the dataset, including duplicates. Use this to measure volume (e.g., number of transactions, number of incidents reported).
  For example, to answer the question "How many orders were placed?"
  the function to use is Count records, applied to the orders field.

• Count distinct values counts the number of unique values in a field. Use this to measure variety (e.g., number of unique cities, number of distinct product categories).
  For example, to answer the question "How many different customers placed an order?"
  the function to use is Count distinct values applied to the customer ID field.

Some charts let you add more than one series so you can compare several values in the same visualization.

To add more than one series, click on the + icon on the right next to the first series. You may add up to 10 separate series.

Once more than one series has been added, stacking options become available for bar and area charts. Three stacking modes are available:

• Separate (default) — Displays each series independently, for when you want to compare series side by side

• Stacked — Combines the series in the same bar or area, for when you want to see both the total and how it breaks down

• Percentage — Shows each series as a share of the total, for when you want to compare proportions, regardless of total size

Use Separate when you want to compare the series individually. Use Stacked when you want to compare both totals and composition. Use Percentage when you want to compare proportions rather than raw values.

The configuration of the Data Explorer (chart type, axes, filters) is automatically reflected in the browser URL. To share your analysis, simply copy and send the URL and the recipient will see the exact same chart you configured.

You can download a chart or an aggregated table with its current filters applied.

1. On the visualization, click the three-dot icon

2. Click the format you want:

• Export as PNG saves the chart as an image. This option is available for charts only.

• Export as CSV, Export as JSON, or Export as Excel downloads the data behind the chart or aggregated table. For an aggregated table, the file contains every row, not only the page you're viewing.

Note that if a chart has more than one series, you download each series as a separate file. The menu lists each format once per series, with the series name in brackets.

You can also embed a single chart or aggregated table into an external page. The embed keeps the visualization's current configuration and applied filters, so what you see is what you share.

1. On the visualization, click the three dot icon.

2. Click Copy embed snippet.

3. Paste the copied <iframe> code into the external page.

Note that this works only for public datasets. If the dataset is private or restricted, visitors to the embedded page see a message saying the content doesn't exist or is restricted, instead of the visualization.

End users can access the Data Explorer on published datasets with chartable tabular data.

To configure how users discover and explore a dataset, refer to the following:

• Configure a dataset's filters

• Configure your asset page

• Defining a dataset schema