Appearance
Visuals
A chart from any set of results — the Results tab, a filtered set, a Group By or a Pivot. Five kinds, and the one you want depends on what you are trying to show.
Where to find it: the Visual tab on a query's results
| What it is for | |
|---|---|
| Bar | Comparing categories — pupils by instrument, income by school |
| Line | A trend, where the categories are a sequence |
| Pie | A share of a whole, and only worth it with few slices |
| KPI Card | One headline number, aggregated from a column |
| Map | Where people are, by postcode district |
A result set with no numbers opens on the Map
Every visual except the Map plots a number, so a query that returns none has nothing the others can draw. Rather than refusing, it opens on the Map — which needs only a post code — so a list of pupils still shows you something.
Bar and line
Pick the category for the X-axis and one or more value columns. Pie and KPI Card take only the first value column, because neither can show more than one series.
| Sort | By category, so a yyyy-MM month reads left to right, or by value, biggest first |
| Sort column | A numeric column that sets the order instead — an assessment level's own Value, say, so levels read in your order rather than alphabetically |
| Series labels | What a value column is called in the legend — Last Year rather than CompareValue |
| One chart per | Splits into a chart for each value of a column, largest first |
A chart shows at most 30 categories
With more than that it keeps 30. Which 30 depends on the sort: sorted by value it keeps the biggest, which is usually what you want, and sorted by category it keeps the first 30 in order.
A saved query's own ORDER BY is dropped when it runs, so with no sort chosen the rows arrive in no particular order — and the 30 you keep are then arbitrary. If a chart looks like it is missing the interesting ones, set the sort.
The Map, and why it is deliberately coarse
The Map counts rows by postcode district — the outward part only, so CF24 1AB counts as CF24 — and draws each district's count at the middle of that district. Pick which column holds the post code where a query returns more than one.
It opens framed on the districts that have pupils, not on the whole country, with a little margin round them — so a service covering one county fills the frame rather than sitting in the corner of an empty island. Fit returns to that view, Whole country zooms out to Great Britain when you want two services at the same scale, and you can zoom with the wheel and drag to pan.
It never plots a person, and that is the point
The rows behind a pupil query are children's home addresses. So the map is aggregate only: a count per district, never a point per row, and nothing finer than a district is ever drawn.
A district holding fewer than three rows shows its position without its number, so a single family in a thinly populated area cannot be read off the map. Use it for "where are our pupils coming from", not for finding anybody.
Rows it could not place are counted, not dropped
Post code coverage across a service's records is usually partial. The map reports how many rows had no post code, and how many had one it does not recognise — a foreign address, or a typo.
Read that number before drawing conclusions. An area that looks empty may be an area whose post codes are missing, and a map that silently dropped them would tell you the opposite of the truth.
Keeping one
A visual can be saved as a dashboard item, printed, or saved as a picture — see working with results for the rest of what a result set can do.
See also
- Group By, Pivot and Charts — the shaping that usually comes first
- Building a query
- Creating and managing dashboards — where a visual ends up living
