Helping you get started
We have produced a series of short videos to help you understand and make best use of the Flexible Table Builder.
Building a simple table
This video shows you how to create a custom table, filter or pivot data within your table, and download your table.
Improving your results
This video shows you a number of techniques that can help improve the results of data confidentiality checks.
Accessing and customising ready-made tables
This video shows you how to access and customise our set of ready-made tables.
User feedback
In our series of short videos, Census Office has endeavoured to cover the basics of the Flexible Table Builder. However, we are always looking to improve users’ experience and we recognise there may be other topics that would benefit from tutorial-style videos.
Please email census@nisra.gov.uk with topic suggestions for additional videos.
Statistical disclosure control approach
There are two main methods used to protect personal data in the Census 2021 results. Both are applied for all aggregate outputs and offer a complementary form of protection.
- Targeted Record Swapping: Used to protect individuals and households with unique or unusual characteristics. This prevents easy spontaneous recognition of individuals and households within datasets.
- Cell Key Perturbation: The cell key method offers protection against disclosure by differencing, where two or more slightly different datasets could be compared to expose an individual respondent, and in instances where a few datasets can be constructed and could otherwise be linked together to reconstruct records from the microdata.
In addition to the above, further disclosure control rules are built into the Flexible Table Builder. The facility to "build a custom table" could potentially allow users to produce an extremely large bank of tables. The combination of these could allow identification of individuals and disclosure of information. Therefore the following disclosure rules have been put in place to limit the detail:
- marginal minimum - where a row or column has a small total, the table can be susceptible to an attribute disclosure, or to help an intruder build up an individual record if that total appears in other tables. In deciding the value of the minimum, we take record swapping and perturbation into consideration and the likelihood of whether the records at risk are real and in the correct geographic area.
- marginal dominance - a variable in a table should not have nearly all respondents in the same category, and there should be at least 20 respondents not in the most common category.
- zeros - data should not contain an excessive proportion of empty cells; at least 40% of the table should be non-zero cells.
- sparsity (ones and zeroes) - a table should not contain an excessive proportion of empty cells and ones. A dataset of ones and zeroes will not only likely be risky, but it also gives a perception of risk, and if less than 50% of the dataset is non-zero cells, at least 50% of the non-zero cells should be larger than 1.
- maximum number of cells - there should be an average of at least one case per cell in the table.
- maximum number of variables - up to four variables can be selected at Data Zone (DZ) or Super Data Zone (SDZ) level, and up to five variables are allowed at Local Government District level and above.
These disclosure checks are automated, rule-based checks run by the system, which decide if there is a low enough disclosure risk to allow the release of a table. The rules allow release for those areas where the risk is sufficiently low, while stopping release in areas where the risk is higher. This "patchwork" approach allows more to be released than the previous "blanket" approach, which would have blocked the table for all areas if some areas were "too risky".
Where a custom table in the flexible table builder is unavailable for some geographic areas because of these disclosure rules, we will not provide details of which rule or rules have caused the areas to fail. Users are advised to consider either reducing the detail for one or more of the variables or using a higher geographic level. Watch our short video on improving your results to get more advice.
For more detail, please see the Statistical Disclosure Control Methodology Guidance note (PDF, 169 KB). Section 4 provides additional guidance when building tables in the Flexible Table Builder.
Revisions and corrections policy
The Flexible Table Builder is a digital product that may routinely be updated with software patches and/or fixes. Any enhancements to the system relating to changes in census data and/or metadata are documented in the Census 2021 quality notes, issues and corrections page.