English and Portuguese names work: gasolina, arroz, milho, câmbio. ↑↓ to move, Enter to open.

Explore relationships

Pick two indicators to see whether they move together, with or without a delay. Useful for questions like “do fuel price rises show up in transport inflation?” For indicators with several markets or countries, the default series is used (Maputo, Mozambique or the national index).

First indicatorPrecious stones (rubies) — exports
Second indicatorInflation (consumer prices, year on year)

Precious stones (rubies) — exports and Inflation (consumer prices, year on year)

What is this?

Two indicators on one chart, aligned on the periods they share. Series of different frequencies are averaged to the coarser one (daily to monthly, monthly to yearly).

How to read it

  • Two scales: each line keeps its own units (left and right axes). Indexed: both start at 100, so you compare percentage movement.
  • Correlation (r) runs from −1 (move in opposite directions) to +1 (move together). Changes-based correlation is less fooled by two series that both simply trend upwards.
  • The shift slider moves the second series in time, to test whether one tends to follow the other.

What does the data say?

Not enough shared periods to calculate a correlation.

Why does it matter?

  • Use it to check whether a suspected driver (fuel, food, the exchange rate) has moved together with an outcome before building a business case on it.
  • A strong correlation is a reason to investigate, not proof: causes must be checked with other evidence, such as internal transaction data.

What should I watch?

Data source

Source
Each indicator's own source (see its page)
Status
Calculated by Stats from the source series
Latest observation
—
Last updated
—
Frequency
Common frequency of the two series
Method
Pearson correlation on aligned levels and on period-to-period percentage changes; best lag searched from −12 to +12 periods; rolling correlation over a fixed window.
Limitations
Correlation does not imply causation. Short, annual or heavily trending series give unreliable results.

Loading…

Correlation measures whether two series tend to move together, from −1 (opposite) through 0 (unrelated) to +1 (together). It does not show that one causes the other: both may respond to a third factor, trends can make unrelated series look correlated, and short or annual series give unreliable results. Series of different frequencies are averaged to the coarser one before comparing.