Data sources

A data source answers one question: where did this number come from? The beneficiary register, the clinic's attendance sheet, a partner's monthly spreadsheet, a household survey. You register it once for a project, and every data collection that uses it points back to the same record.

Why it matters

Months after a project ends, a donor or an auditor asks where a figure came from. With your sources registered, the answer is one click away:

indicator → monitoring plan → data collection → data source

Without them, the number is just a claim. The data source is the link people skip most often, because a data collection saves without one.

Add a data source

Open Data Sources and click to create a new one:

Field What to enter
Project The project that uses it. Only active projects are listed
Name Something a stranger could find: the institution and the document, such as "Clinic A attendance register", not "monthly numbers"
Type What the source is (see below)
Description Who holds it, how often it is updated, what it leaves out, and who to ask

Tip The description is the most useful field and the one most often left empty. There is no separate field for the owner or the update frequency, so write both here.

The types

Type Means
Register An existing list you read from
Register form A form filled in as part of daily work
Register review Going back through records already kept
Survey A structured questionnaire
Interview A structured or semi-structured conversation
Observation Someone watching and recording directly
Excel A spreadsheet handed over by someone
API Another system sending figures automatically

Source is not method

These are two different questions:

  • Source: what you read, such as the district health register.
  • Method: how you got the figure out of it, such as a register review. The method belongs on the monitoring plan, not here.

Register a separate source whenever the thing being read is different. "Clinic A register" and "Clinic B register" are two sources, even if you read them the same way.

Keep the list healthy

  • Each source shows how many data collections used it. A source used many times is one your indicators depend on: if it disappears, record that in the risk register.
  • A source used zero times after a few months was either never needed or never used. Start using it or remove it.
  • Look for near-duplicate names. That is usually the same register entered twice.

Warning Deleting a source does not delete the data collections that used it. They simply stop naming a source. To retire a source that has been used, say so in its description instead of deleting it.

What's next

Record your data in data collections, and choose the source each time.