### Relocation-class indicators describe long-term (weekly, monthly, seasonal) changes in the number of people changing their home location between pairs of areas.

You can read more about our methods for assigning home locations here.

## Relocations

The relocations indicator estimates the number of people who relocate (i.e. change their home location) from area A to area B during a given period of time.

The usual resolution for this indicator is:

• Temporal: monthly

This indicator describes the numbers of people changing where they are living or staying between pairs of areas within a country. This allows us to understand how the numbers of people relocating between areas varies over time.

Calculation

We calculate the relocations indicator using the home relocations aggregate from the directed pairwise number of home relocations between areas.

For a given pair of areas, we estimate the number of subscribers who are relocating from area A to area B as the unredacted count of subscribers whose home location changed from area A to area B during a given period of time.

We adjust the number of relocations with respect to the changes in the number of active subscribers in the destination area. The adjustment factor is calculated as the median number of subscribers who continue to reside in the area between time steps during the baseline period divided by the number of subscribers who continued to reside in the area during the period of interest.

We can also convert the indicator from the number of subscribers relocating to the number of people relocating using a static population figure, such as that produced by a census. We may also adjust the indicator for representation biases based on mobility errors estimated from survey data.

## Relocations difference

The relocations difference indicator estimates the difference in number of relocations from area A to area B between the time period of interest and a reference period. If the number of relocations has increased, the indicator will have a positive value; if the number has decreased, the value will be negative.

The usual resolution for this indicator is:

• Temporal: monthly

This indicator describes the change in the number of relocations between pairs of areas, relative to a baseline number of relocations during a reference period.

Calculation

We calculate the relocation difference indicator using the home relocations aggregate and the baseline number of relocations between the areas during a reference period. The baseline can be calculated using the home relocations aggregate, for example by calculating the median number of relocations during the reference period.

We estimate the change in the number of relocation between pairs of locations by deriving the number of relocations for the period of interest from the home relocations aggregate as described above. We then subtract the baseline number of relocations from the number of relocations during the period of time of interest.

We adjust the number of relocations with respect to the changes in the number of active subscribers in the destination area as described above.

We can also convert the indicator for the change in the number of subscribers relocating to the number of people relocating using a static population figure, such as that produced by a census. We may also adjust the indicator for representation biases based on mobility errors estimated from survey data.

## Percent change in relocations difference

The percentage change in relocations difference indicator estimates the change in relocations between area A to area B between the time period of interest and a reference period, relative to the number of relocations between these areas during the reference period. The difference is expressed as a percentage of a baseline value calculated for the reference period. If the number of relocations has increased, the indicator will have a positive value; if the number has decreased, the value will be negative.

The usual resolution for this indicator is:

• Temporal: monthly

This indicator describes the change in the number of relocations between pairs of areas, relative to a baseline number of relocations during a reference period.

Calculation

We calculate the difference in the number of residents relative to the baseline as described above. We divide this difference by the baseline number of residents  in the reference period and multiply by 100 to derive the percentage change.

We adjust the number of relocations with respect to the changes in the number of active subscribers in the destination area as described above.

We can also convert the indicator for the change in the number of subscribers relocating to the number of people relocating using a static population figure, such as that produced by a census. We may also adjust the indicator for representation biases based on mobility errors estimated from survey data.

## Abnormality

The relocations abnormality indicator measures the deviation of the number of relocating from area A to area B during the time period of interest relative to a reference period, expressed as a z-score. A positive value for this indicator greater than 3 indicates a statistically significant increase in the number of relocations; a value less than -3 indicates a statistically significant decrease in relocations.

The usual resolution for this indicator is:

• Temporal: monthly

This indicator describes how unusual the number of relocations from area A to area B is, given the amount of variation observed during the reference period. Higher absolute values indicate greater deviation from the normal variation in relocations, and therefore greater probability that the change is meaningful. Such changes may be associated with specific events or may be caused by technical issues. Very large abnormalities (absolute values greater than 6) may, however, be more indicative of a technical issue.

Calculation

We calculate the relocations abnormality indicator using the home relocations aggregate and the baseline number of residents in the area during a reference period. The baseline can be calculated using the home relocations aggregate, for example by calculating the median number of relocations from area A to area B during the reference period.

The relocations abnormality is the z-score, or number of standard deviations the number of relocations in a given period is above or below the median number of residents in the reference period.

To calculate the z-score, we calculate the difference in the number of relocations relative to the baseline as described above. We then divide this difference by the median absolute deviation (MAD) in the number of relocations during the reference period.