Chart Kind Gallery

This is the encyclopedia entry for every chart kind BankSync can draw. The Chart Kind Catalog groups kinds by the question they answer and helps you choose one. This page goes one level deeper: for each kind it tells you what it shows, which fields it needs (its channels), and what it is good for. Use it when a kind is not behaving the way you expect, or when you want to know exactly what data a kind wants before you bind it.

A channel is a slot a chart binds a field into. A bar chart, for example, has an X-axis channel (the category) and a Y-axis channel (the number). Different kinds have different channels, which is why the controls in the inspector change when you switch kinds. The rest of this page is organized by family: kinds in the same family share the same channels and behave the same way, so once you learn one kind in a family you know them all.

How to read this page#

Each family has a short table. The columns are:

• Kind: the name you see in the picker.
• What it shows: the picture it draws.
• Needs: the fields you must bind (the required channels). Optional channels are noted in the family intro, not the table.
• Good for: the one situation it is built for.

A worked example: reading one row#

Say you bind a sankey (in the Flow family). Its row reads:

• Needs: Source (a category), Target (a category), Value (a number).
• What it shows: ribbons flowing from each source to each target, with ribbon thickness set by the value.
• Good for: showing how money moves between accounts or from income into spending categories.

So to build "money flow between accounts" you bind your from-account field to Source, your to-account field to Target, and the amount to Value. The chart sums the amount for every repeated from/to pair, so duplicate transfers between the same two accounts merge into one thick ribbon. That summing behavior is covered in Does this kind add up my rows? at the end.

Cartesian: lines, bars, and areas on an x/y plane#

The largest family. Every kind here plots a category or date on the X axis and a number on the Y axis, with two optional extra slots: Series (split one line or bar into several, one per value) and Color. These are the everyday workhorses.

Scatter: relationships between two numbers#

Both axes here are numbers, not categories. Two optional slots add a third and fourth dimension: Size (a bubble) and Color.

Distribution: how values are spread out#

These take a Value (a number) and optionally a Category to split by. They summarize the spread internally, so you do not bind a separate total.

Histogram: how often values fall in each range#

Bin a single number and count how many rows land in each bin. There is one channel: the value to bin.

Part of a whole: slices of a total#

These need a Category to slice by and a Value for each slice's size. Pie and Donut are the everyday choices; the rest are stylistic.

Hierarchical: nested shares#

These read a Category and a Value and nest the categories into levels. The renderer reshapes your fields into a name/value/children structure for you.

Flow: how amounts move between things#

These bind two categories, Source and Target, plus a Value for the magnitude. None of them use an axis.

Matrix: intensity across two dimensions#

These bind an X category, a Y category, and a Value shown as color or dot size in a grid.

Calendar: a value per day#

One channel for the Date and one for the Value, laid out as a calendar.

Radar: measures around a circle#

Binds an Axis category (one spoke per value), a Value, and an optional Series (one polygon per series).

Multi-dimension: many measures at once#

These read several numeric columns at once rather than one bound value. They expose only an optional Group color slot; the measures are read from the data's numeric columns (or from a column list you shape in the query).

Financial: price movement#

These plot price over time. They bind a Date axis, and they read the price columns by fixed name: Open, High, Low, Close (named o/h/l/c). You shape those columns in the query rather than binding them as channels.

Single value: one number on a scale#

These show a single number, so they have just one Value channel and no axes. They are the simplest kinds to bind.

The single-number widgets (Gauge, Meter, Progress) overlap with the KPI widget, which also shows one figure but adds comparisons, sparklines, and number formatting. For a headline number with context, see KPI, table & gauge widgets.

What about column, KPI, and table?#

A few kinds you will use a lot are not in the families above:

• Column (vertical bars) is the bar family drawn upright; bind it exactly like a bar.
• KPI and Table are not chart kinds at all, they are their own widget types. KPI shows one headline number with comparison and a sparkline; Table shows formatted rows. Both have their own page: KPI, table & gauge widgets.

Does this kind add up my rows?#

Some kinds total your rows before plotting; others plot every row as-is. This is why a Sum setting changes a bar chart but does nothing on a histogram.

For the kinds that total rows, the Binding data page covers the aggregate control (Sum, Average, Count, and so on) and the double-count warning that appears if your data was already totaled in the Shape data step. For the kinds that plot rows as-is, an aggregate setting has no effect, so you will not see one.

The less-common families (Flow, Hierarchical, Matrix) render exactly like the everyday ones once bound: a sankey or treemap sits in a normal widget cell on the canvas, draws from your live feed data, and resizes with the grid like any other chart.

Switching kinds keeps your fields#

Because kinds in the same family share channels, switching between them is a one-click change that keeps your bound fields. Line to area to column keeps the date and the number. Switching across families can drop fields the new kind has no slot for: a multi-series line moved to a pie (which has only a category and a value) loses its series split. Switch back and you re-bind it. See The Chart Kind Catalog for more on switching.

Related guides#

• The Chart Kind Catalog: choose a kind by the question it answers.
• Binding data: set the fields each kind needs and how to total them.
• KPI, table & gauge widgets: the non-chart widget types.
• Adding widgets: get a widget of any kind onto the canvas.