Skip to main content

Data Privacy

Overview​

As the data visualisation feature is accessible to all logged-in users, data is aggregated and suppressed to reduce the risk of re-identification and revealing raw values. This page describes the aggregation and suppression methodology for the following consumption datasets: EMA Electricity, EMA Towngas and PUB Water.

Aggregation & suppression​

The data is aggregated monthly at the postal code level for each dataset. The median number of accounts at each postal code is computed for each dataset, and used to compute the average consumption and account number displayed.

info

If a postal code is marked as invalid, it is not in the latest SLA Address Point dataset. The postal code might no longer be in use if it is in your mapping file.

If a postal code is marked as suppressed, the max(account no) over all months is ≤ the suppression level.

To make more data available for querying, some postal codes are also grouped into residential or industrial estates. If a desired postal code is within an estate, they cannot be queried individually.

Here is an example of an estate in the dataset:

Estate Highlighted

Suppression level​

The suppression level (n) is 4. This means that any postal code or estate with ≤ 4 accounts over all historical months for the dataset will not be shown in the data visualisation feature.

This number was selected using an elbow chart, and picked as there are diminishing marginal returns (in terms of proportion of postal codes suppressed) from further increasing n. The same suppression level is applied on all 3 datasets.

Get data for suppressed postal codes​

You can either submit a data request or contact us at support@datahive.gov.sg about adding an estate to the feature.