Step by Step Guide: Building Pivot Tables and Charts, Using ANC Visits as the Example
Data that has been entered is of no use until somebody looks at it. This guide builds a pivot table and then every kind of chart from the same selection, so you can see what each one is good for. It uses the four ANC visit data elements, because they hold data in every demo, and the steps are identical for your own malaria data elements and indicators. The charts you save here are what the next topic puts on a dashboard.
Charts and pivot tables are built in the Data Visualizer. Open the apps menu and pick it from the list, or type the first letters of the name.
The app opens on an empty pivot table. Everything you build is made of three things: the type of visualisation, the dimensions you pick, and where you put them.
Click Data in the left list. Choose the type on the left of the window, for example Data elements or Indicators, then move the items you want to the right hand list and click Update. This guide uses the four ANC visit data elements as its example, because they exist in every demo.
Click Organisation unit. You can tick units in the tree, or choose a whole level or group, which is usually what you want for a chart.
Click Period. Relative periods move with time, which is what a dashboard needs, and fixed periods stay where they are, which suits a report about one month.
With data, period and organisation unit chosen, click Update. The pivot table appears, and it is worth reading before drawing any chart, because it shows exactly the numbers the chart will use.
Open the visualisation type menu at the top left. Each type suits a different question, and the description under each name says what it is for.
A line chart answers how something has moved month by month. An area chart does the same but stacks the series, so the height of the whole shape is the total.
Columns stand up and bars lie down. They carry the same information, so the choice is about the labels: long names read better on a bar chart.
Stacked charts show the parts and the total at the same time. A pie chart shows the shares at one point in time, so the period usually goes in the filter.
Year over year puts this year against last year on the same twelve months, which is the quickest way to see whether a season is worse than usual.
| The question | The chart |
|---|---|
| What are the numbers | Pivot table |
| How has it moved over the months | Line |
| How do the parts add up over time | Area or stacked column |
| Which is bigger | Column, or bar for long names |
| What share does each have | Pie |
| Is this year worse than last year | Year over year |
| Where is it happening | Open as Map |
A chart has to be saved before it can be put on a dashboard. Give it a name that says what it shows, including the period and the level, so it can be found later.
| Good name | Malaria: confirmed cases by district, last 12 months |
|---|---|
| Poor name | Chart 1 |