The charts (maps, graphs, plots, diagrams and tables) available in **ArcGIS Insights** to visualize data and the stories within, can be used for measure and comparison; to view distribution or arrangement; or see the connections and similarities between categories or objects.

Here I’ll offer a breakdown of each chart type, and strive to give an insider’s opinion on what the intention of each chart is, and (through the links) some advice on when and how to use them. *Skip to the end for a workbook that contains all chart types.*

### Location Map

To indicate where places of interest are, or to show the geographical coverage and spread of a dataset.

Click here for tips and an example

### Graduated Symbol Map

Also known as a dot map, a graduated symbol map uses symbol size to represent quantitative (numeric) data based on location.

Use these to compare statistics at different geographic locations, e.g. the number of reported infections by town or city.

Proportional symbol maps are a specific case where data is unclassed, and the symbol area directly relates to the numeric value being shown.

Click here for tips and an example

### Choropleth Map

Available when styling polygons by “Counts and amounts”, a choropleth map uses color to represent quantitative data tied to geographic areas, usually a seamless set of familiar polygons such as US States.

Used to show rates, ratios or percentages (e.g. number of responders per square mile).

Click here for tips and an example

### Binned Map and Heat Map

In Insights we have several ways to represent density on a map.

A binned map converts point data to a grid, with each grid square (or bin) colored by the summary statistic of all the features within. For example, if a bin contains 5 data points, we can take a numeric field and color that bin by the average value or the summed value of the five entries.

Heat maps interpolate an estimated surface density from point data. They are a quick and visually powerful way to show “hot spots” and spatial distribution.

Click here for tips and an example

### Pie or Column Chart Map

A pie chart map is a map card with pie charts displaying the distribution of counts or numeric values between different categories. Each pie is associated with a specific geographic extent or location.

Pies are ideal for giving a quick idea of the proportional distribution which can be further examined via the hover tooltip. The pies themselves vary in size based on the totals associated with each geography.

Columns can be stacked on top of one another or expanded to be side-by-side. Comparisons can be made using either counts or percentages.

Click here for tips and an example

### Bar Chart and Column Chart

A bar chart is used to compare values (either counts or from a numeric field) between categories. Each category is listed along the axis and the width of each bar (extending out to the right) is proportional to the value it represents.

A column chart is a specific case where the bars extend vertically upwards and labelling options update accordingly.

Click here for tips and an example

### Donut Chart

A donut chart shows each category as a proportion of the total value, like a pie chart. The hole in the center is used to display the total amount. The values can either be a count or come from a numeric field of the dataset.

Donuts are ideal for showing proportional distribution between categories.

Click here for tips and an example

### Histogram

Histograms sort all the values from a number field into bins. The height of each column tells the frequency of each bin. This shows the distribution of the numeric variable.

The bins are aggregated classes of a numeric field, such as age groups or time elapsed, and as such are continuous. The user can control the number of bins shown.

The frequency is a summed count of the number of times the value defined by the bin occurs in the dataset.

*Traditionally the class width of a histogram’s bins can vary and so the frequency count of each bin is represented by the area of the columns rather than their height. This is generally only needed if your source data is already binned, and it is not supported in Insights.*

Click here for tips and an example

### Scatter Plot

A scatterplot shows a point for each record of data, representing a pair of values from two numeric fields. It provides an organized display of the data which helps show patterns, distribution, outliers and primarily the strength of any relationship between the two numeric variables.

With chart statistics active, a trendline is shown to approximate the data, along with its equation and associated R-squared value – measuring how well the trendline fits the data and indicating how much of the variance seen in one variable is potentially explained by the other.

Click here for tips and an example

### Time Series

A time series is a form of line graph where the x-axis is date-time. It can be used to visualize trends in counts or numerical values over time.

As date-time data often consists of a very large number of records, Insights intelligently amalgamates data into “bins”. You will see the bin ranges on the tooltip when hovering over the chart.

Click here for tips and an example

### Treemap

Treemaps are a set of nested rectangles or blocks each representing a different category. These are sized by values from one numeric field and can optionally be colored by values from another. They can also be grouped by a second category (known in Insights as subgroup).

The rectangles are arranged hierarchically, firstly into clusters of the same subgroup ordered by their total group values, and secondly by their individual category sub-totals, with the greatest values in the top-left and the smallest bottom-right.

They are a compact way to show a lot of information and the tree-like structure can help spot patterns missed by other chart types.

Click here for tips and an example

### Bubble Chart

Each bubble represents a category from a string (text) field. The size of these nested circles is proportional to either their count or the value from a chosen numeric field. The positioning of the bubbles is not significant but is optimized for compactness.

Quick visual comparison of the size of bubbles is possible due to their close proximity. However, the chart is also a great way to show large outliers.

*Our implementation is also known as a packed bubble chart. For scatter bubble charts please see Scatter Plot.*

Click here for tips and an example

### Line Graph

Line graphs or line charts are used to show values when there is a sequence to the categories on the x-axis. They are formed of a series of data points, which represent a category and associated count or numeric value, connected by lines.

They are also a useful alternative to column charts, particularly when showing smaller variations.

*Some people use the term line graph or chart to mean values plotted against time, for that please see *Time Series*. We can use time components such as “Year” or “Day of Month” on a line graph though.*

Click here for tips and an example

### Chord Diagram

A chord diagram is a visual matrix. Categories are arranged in a circle of outer arcs, from which relationships or flows between categories are illustrated by connecting chords. Values associated with each connection or category pair, are represented proportionally by the size of each chord.

They are ideal for establishing relationships or identifying dominant categories by comparing the similarity or differences between pairs or volume of flow.

Click here for tips and an example

### Data Clock

A way to visually bin data, our data clock is mainly used to show distribution between categories summarized over a standard time period such as hours of the day or days of the week.

The data clock is a series of concentric circles or rings divided into wedges. The concentric rings represent different categories and the wedges represent a second category of breakdown, usually a component of date-time as described above. The individual cells or bins are color coded in similar fashion to a heat chart with darker colors usually representing higher values.

Click here for tips and an example

### Heat Chart

Also known as heat maps, these grids of colored cells allow us to view multivariate data through placing categories from one variable in the rows, and from another in the columns. The cell is colored by the numeric value associated with the two variables in the connecting row and column – a bit like coloring in a spreadsheet by the values in the cells. This number can be the count or the value from a numeric field.

They are useful in spotting hierarchies, cluster groups (cells or bins of similar value), and outliers.

Click here for tips and an example

### Box Plot

Box (and whisker) plots summarize the variance or data distribution between values in each category. At a glance we can see where the bulk of the data lies, on inspection we can see a lot more…

Horizontal lines represent the median and extreme values, the range between upper and lower quartiles (or that of the central 50% of data) is represented by a column called a box, and any outliers are shown as points.

Extreme values are the lines at the end of the “whiskers” and are the minimum and maximum values within 1.5 times the inter-quartile range.

Box plots have the ability to summarize a lot about the data while taking up little space on the page. They allow us to compare multiple categories or groups, and for each can tell us what the key values are, if there are any outliers, if the data is skewed or symmetrical, and how tightly it is grouped or how much it is spread.

Click here for tips and an example

### Link Chart

Link charts are a form of network diagram used to show the magnitude and direction of relationships between categorical data. They’re used in link analysis for identifying relationships or links between nodes that are hard to see in traditional chart types or tables.

The nodes often represent items such as people, places or events and are sized by centrality – a measure of importance or connectivity.

They expose hierarchical relationships, and common links highlight influential nodes, for example a good place to focus any outside influence such as advertising. They can more generally be used to analyze the structure of a network.

Click here for tips and an example

### Alluvial Diagram

Alluvial diagrams are a type of flow diagram showing changes in group composition between category fields. The thickness of each stream or link shows its proportional value.

It is easy to see what is connected, and where the total is spread. Use one to identify relative strength of links, or trace multiple paths through a hierarchy.

Click here for tips and an example

### Scatterplot Matrix

A scatter plot matrix is a grid of scatter plots, one for each combination (pair) of up to five numeric variables. It quickly shows which variables have the strongest correlation or relationship, as well as any outliers.

It also shows the R² value for each plot, an indicator of how much of the variance seen in one variable is potentially explained by the other, with 1 being all and 0 being none. The matrix is also a good starting point for regression analysis.

Click here for tips and an example

### Key Performance Indicator (KPI)

A KPI is a metric used to compare the status or value of a measure based on data at hand against a key goal or target.

They are often used for performance benchmarking, revenue improvement, cost reduction, process efficiency, and customer satisfaction.

On our card, we show the measure (so the current or chosen value) in large numbering with the target shown beneath. Through the layer options an accompanying gauge visual can be added.

Click here for tips and an example

### Combo Chart

A “combo” chart is a combination of two column charts, two line graphs, or a column chart and a line graph. You can make a combo chart with a single dataset or with two datasets that share a common string field.

Comparing values from two different numerical variables based on common categories, either side-by-side as columns, or overlaid as a column and a line, gives a clear view of which measure is higher or lower and the difference between the two. It is also a clear way to show the difference in distribution between the two measures.

Click here for tips and an example

### Point Chart

A point chart is a bit like a bar chart but with a point for each value associated with each category. It highlights typical values of each category as well as visualizing the whole range and distribution of the data, in particular range and variation.

For single value categories, we can add confidence intervals or error bars too. Point charts can be created on their own or as part of a regression analysis.

Click here for tips and an example

### Summary Table

A summary table groups data by category and the aggregates the associated values into summary statistics such as sum, minimum, maximum, average, median and percentile.

Summaries for each column, for example totals, are shown at the bottom.

Click here for tips and an example

### Reference Table

A reference table shows the data table “raw” without any aggregation. So, as it is stored within Insights or as it has been read into the analytics workbook.

The difference between this and viewing the data table, is you can pick and choose which fields to show, and it produced a card on the page.

In addition, reference table has table and column formatting, a range of conditional formatting options – including rule based and data bars, sparklines – including columns and win/loss, and it can contain clickable web links.

Click here for tips and an example

## Article Discussion: