Anonymization Techniques
Learn the IT-English vocabulary of data anonymisation: pseudonymisation, masking, aggregation, k-anonymity and re-identification.
0 / 22 completed
1 / 22
What is the difference between 'anonymisation' and 'pseudonymisation'?
Pseudonymised data can be re-linked via a key; truly anonymised data cannot be traced back to a person.
2 / 22
'Masking' a credit card to show only the last four digits is an example of what?
Masking hides part of a field (e.g. ****1234) so the sensitive portion isn't displayed.
3 / 22
What is 're-identification'?
Re-identification is the risk that supposedly anonymous data can be linked back to individuals.
4 / 22
A dataset satisfies 'k-anonymity' with k=5. What does that ensure?
k-anonymity means any individual blends into a group of at least k people on identifying attributes.
5 / 22
Which sentence correctly uses 'aggregation' for privacy?
Aggregation reports group-level statistics, protecting individuals by not exposing single records.
6 / 22
Reviewer: 'I'm concerned about this user ID field in the `user_profiles` table. It's directly tied to a real person and could be exposed if we ever have a data breach. We absolutely *must* implement some form of pseudonymization before any further development.
You: I understand the risk, but I was planning on just hashing it using SHA-256. It's relatively standard and will provide a good level of protection. What's the best way to respond effectively?This scenario highlights the importance of understanding that simply hashing data isn't always sufficient anonymization. While SHA-256 provides cryptographic security, it doesn't prevent *re-identification* if other attributes in the dataset are known. The reviewer is correct to point out that a breach could expose this directly linked user ID. Pseudonymization, as the reviewer suggests, involves replacing identifiable data with a unique code, which offers greater protection against future disclosure and re-identification attempts.
7 / 22
Reviewer: 'We need to implement differential privacy into our logging system. Specifically, we should add Laplace noise to any metric that's being tracked – things like page views, user session durations, and click counts. This will ensure that no single individual's data can be traced back through the aggregated logs.' What is the primary goal of this approach?Differential privacy aims to protect individual privacy within a statistical analysis. Adding Laplace noise, as described, effectively obscures the contribution of any single record to the overall aggregate metrics. This prevents adversaries from inferring information about specific individuals by observing the noisy data; it's not primarily focused on outlier reduction or performance improvements, nor does it directly address GDPR compliance (though it contributes positively). The core principle is that an adversary cannot reliably determine whether a particular individual's data was included in the dataset.
8 / 22
PR Description:
"Implemented new user profile API endpoint. Includes retrieval of full name, email, and user ID."
During a code review, your colleague comments: 'I'm noticing we're exposing the raw user_id in this API response. While it's technically immutable through hashing, it still presents a significant re-identification risk if combined with other data points. What's the most appropriate way to address this concern during the PR discussion?'The correct answer highlights the critical issue of re-identification, even with hashing. While SHA-256 provides one layer of obfuscation, combining this hashed value with other available information (like timestamps or IP addresses) can potentially lead to a user being identified. Tokenization offers a stronger defense by replacing the original identifier with an entirely new, non-sensitive value – preventing any link back to the individual user. Options A and C are incorrect because they either ignore the re-identification risk or rely on a single layer of security that isn't sufficient.
9 / 22
Reviewer: 'I'm concerned about this user ID field in the `user_profiles` table. It's directly tied to a real person and could be exposed if we ever have a data breach. We absolutely *must* implement some form of pseudonymization before any further development.
You: I understand the risk, but I was planning on just hashing it using SHA-256. It's relatively standard and will provide a good level of protection. What's the best way to respond effectively?This scenario highlights the importance of understanding that simply hashing data isn't always sufficient anonymization. While SHA-256 provides cryptographic security, it doesn't prevent *re-identification* if other attributes in the dataset are known. The reviewer is correct to point out that a breach could expose this directly linked user ID. Pseudonymization, as the reviewer suggests, involves replacing identifiable data with a unique code, which offers greater protection against future disclosure and re-identification attempts.
10 / 22
Reviewer: 'We need to implement differential privacy into our logging system. Specifically, we should add Laplace noise to any metric that's being tracked – things like page views, user session durations, and click counts. This will ensure that no single individual's data can be traced back through the aggregated logs.' What is the primary goal of this approach?Differential privacy aims to protect individual privacy within a statistical analysis. Adding Laplace noise, as described, effectively obscures the contribution of any single record to the overall aggregate metrics. This prevents adversaries from inferring information about specific individuals by observing the noisy data; it's not primarily focused on outlier reduction or performance improvements, nor does it directly address GDPR compliance (though it contributes positively). The core principle is that an adversary cannot reliably determine whether a particular individual's data was included in the dataset.
11 / 22
PR Description:
"Implemented new user profile API endpoint. Includes retrieval of full name, email, and user ID."
During a code review, your colleague comments: 'I'm noticing we're exposing the raw user_id in this API response. While it's technically immutable through hashing, it still presents a significant re-identification risk if combined with other data points. What's the most appropriate way to address this concern during the PR discussion?'The correct answer highlights the critical issue of re-identification, even with hashing. While SHA-256 provides one layer of obfuscation, combining this hashed value with other available information (like timestamps or IP addresses) can potentially lead to a user being identified. Tokenization offers a stronger defense by replacing the original identifier with an entirely new, non-sensitive value – preventing any link back to the individual user. Options A and C are incorrect because they either ignore the re-identification risk or rely on a single layer of security that isn't sufficient.
12 / 22
Reviewer: 'I'm concerned about this user ID field in the `user_profiles` table. It's directly tied to a real person and could be exposed if we ever have a data breach. We absolutely *must* implement some form of pseudonymization before any further development.
You: I understand the risk, but I was planning on just hashing it using SHA-256. It's relatively standard and will provide a good level of protection. What's the best way to respond effectively?This scenario highlights the importance of understanding that simply hashing data isn't always sufficient anonymization. While SHA-256 provides cryptographic security, it doesn't prevent *re-identification* if other attributes in the dataset are known. The reviewer is correct to point out that a breach could expose this directly linked user ID. Pseudonymization, as the reviewer suggests, involves replacing identifiable data with a unique code, which offers greater protection against future disclosure and re-identification attempts.
13 / 22
Reviewer: 'We need to implement differential privacy into our logging system. Specifically, we should add Laplace noise to any metric that's being tracked – things like page views, user session durations, and click counts. This will ensure that no single individual's data can be traced back through the aggregated logs.' What is the primary goal of this approach?Differential privacy aims to protect individual privacy within a statistical analysis. Adding Laplace noise, as described, effectively obscures the contribution of any single record to the overall aggregate metrics. This prevents adversaries from inferring information about specific individuals by observing the noisy data; it's not primarily focused on outlier reduction or performance improvements, nor does it directly address GDPR compliance (though it contributes positively). The core principle is that an adversary cannot reliably determine whether a particular individual's data was included in the dataset.
14 / 22
PR Description:
"Implemented new user profile API endpoint. Includes retrieval of full name, email, and user ID."
During a code review, your colleague comments: 'I'm noticing we're exposing the raw user_id in this API response. While it's technically immutable through hashing, it still presents a significant re-identification risk if combined with other data points. What's the most appropriate way to address this concern during the PR discussion?'The correct answer highlights the critical issue of re-identification, even with hashing. While SHA-256 provides one layer of obfuscation, combining this hashed value with other available information (like timestamps or IP addresses) can potentially lead to a user being identified. Tokenization offers a stronger defense by replacing the original identifier with an entirely new, non-sensitive value – preventing any link back to the individual user. Options A and C are incorrect because they either ignore the re-identification risk or rely on a single layer of security that isn't sufficient.
15 / 22
Reviewer: 'I'm concerned about this user ID field in the `user_profiles` table. It's directly tied to a real person and could be exposed if we ever have a data breach. We absolutely *must* implement some form of pseudonymization before any further development.
You: I understand the risk, but I was planning on just hashing it using SHA-256. It's relatively standard and will provide a good level of protection. What's the best way to respond effectively?This scenario highlights the importance of understanding that simply hashing data isn't always sufficient anonymization. While SHA-256 provides cryptographic security, it doesn't prevent *re-identification* if other attributes in the dataset are known. The reviewer is correct to point out that a breach could expose this directly linked user ID. Pseudonymization, as the reviewer suggests, involves replacing identifiable data with a unique code, which offers greater protection against future disclosure and re-identification attempts.
16 / 22
Reviewer: 'We need to implement differential privacy into our logging system. Specifically, we should add Laplace noise to any metric that's being tracked – things like page views, user session durations, and click counts. This will ensure that no single individual's data can be traced back through the aggregated logs.' What is the primary goal of this approach?Differential privacy aims to protect individual privacy within a statistical analysis. Adding Laplace noise, as described, effectively obscures the contribution of any single record to the overall aggregate metrics. This prevents adversaries from inferring information about specific individuals by observing the noisy data; it's not primarily focused on outlier reduction or performance improvements, nor does it directly address GDPR compliance (though it contributes positively). The core principle is that an adversary cannot reliably determine whether a particular individual's data was included in the dataset.
17 / 22
PR Description:
"Implemented new user profile API endpoint. Includes retrieval of full name, email, and user ID."
During a code review, your colleague comments: 'I'm noticing we're exposing the raw user_id in this API response. While it's technically immutable through hashing, it still presents a significant re-identification risk if combined with other data points. What's the most appropriate way to address this concern during the PR discussion?'The correct answer highlights the critical issue of re-identification, even with hashing. While SHA-256 provides one layer of obfuscation, combining this hashed value with other available information (like timestamps or IP addresses) can potentially lead to a user being identified. Tokenization offers a stronger defense by replacing the original identifier with an entirely new, non-sensitive value – preventing any link back to the individual user. Options A and C are incorrect because they either ignore the re-identification risk or rely on a single layer of security that isn't sufficient.
18 / 22
Reviewer: 'This API endpoint returns the full name and user ID of every user. We need to ensure we aren't inadvertently revealing personally identifiable information (PII). What's the most appropriate action in this situation?
Option A: Simply add a comment to the code explaining why we're returning the user ID.
Option B: Implement pseudonymization by replacing the user IDs with randomly generated, unique identifiers while still allowing retrieval of full names.
Option C: Aggregate the data to remove individual user IDs before reporting it, perhaps grouping users based on location and activity level.
Option D: Immediately flag this as a high-severity risk and request a complete redesign of the API endpoint.
Aggregating data is a common technique for reducing PII exposure. By grouping users based on certain attributes, you can obscure individual identifiers while still retaining valuable analytical information. Options A & B are insufficient because they don't address the fundamental risk of exposing raw user IDs; option D is overly aggressive without understanding the broader context.
19 / 22
Sarah (Lead Data Engineer): 'I'm getting reports that our new customer segmentation model is generating highly specific profiles based on seemingly innocuous data like browsing history. We need to ensure we're not inadvertently re-identifying customers through this process. What's the best way to communicate this concern to the team?
Option A: Send a detailed technical document outlining the potential risks of re-identification.
Option B: Start a Slack thread asking for suggestions on how to mitigate the risk, encouraging brainstorming and open discussion.
Option C: Schedule a formal meeting with the data science team to present the findings and request immediate action.
Option D: Post an urgent alert in the #data-alerts channel notifying everyone of the potential issue.
A Slack thread facilitates collaborative problem-solving and allows for diverse perspectives. While a formal meeting or technical document has its place, a Slack discussion is more agile and likely to yield quicker solutions. Option A is too dense; Option C is overly bureaucratic; and option D is an inefficient way to initiate a complex conversation.
20 / 22
PR Description:
"Implemented new user profile API endpoint. Includes retrieval of full name, email, and user ID."
During a code review, your colleague comments: 'While this API fulfills the immediate requirements, it's exposing the raw user_id which could be used to link back to individual users if combined with other data sources. We need to consider how we can mitigate this risk.' Which of the following is the *most* appropriate next step?
Option A: Remove the full name and email fields from the API response.
Option B: Implement a hashing algorithm on the user ID before returning it.
Option C: Add a disclaimer stating that the user ID may be used to identify users in conjunction with other data.
Option D: Immediately revert the changes and request a complete redesign of the API endpoint.
Adding a disclaimer is a pragmatic step that acknowledges the risk without fundamentally altering the functionality. Hashing (Option B) doesn't eliminate the risk entirely; removing fields (Option A) might not be feasible or desirable; and reverting immediately (Option D) is overly reactive without exploring alternative solutions. The disclaimer provides transparency and context.
21 / 22
David (Backend Developer): 'Today I've been working on the new user profile API. It currently returns the user's full name, email address, and their unique user ID. I was discussing privacy with Maria and she raised concerns about potential re-identification risks. How should I frame this during my daily standup?
Option A: 'I'm building a new API endpoint for user profiles.'
Option B: 'I've added a field to the user profile API that returns the raw user ID – Maria flagged some privacy concerns.'
Option C: 'I'm resolving an issue with data aggregation in the user profile service.'
Option D: 'I'm optimizing the performance of the user profile API.'
Option B is crucial because it directly addresses Maria's concern and highlights the potential risk. It doesn't shy away from the technical details but frames them within the context of a privacy issue. The other options are irrelevant to the core problem.
22 / 22
The following is an API response returned by our customer support system:
```json
{
"ticket_id": "CT-20231027-001",
"customer_name": "John Doe",
"email": "john.doe@example.com",
"user_id": "USR12345",
"request_timestamp": 1698400000
}
```
During a code review, your colleague points out: 'We're returning the `user_id` directly. This could be used to re-identify the customer if combined with other information. How can we improve this response to minimize privacy risks?'
Option A: Add a warning message to the API documentation advising developers to avoid using the `user_id`.
Option B: Implement a system that generates a unique, random identifier instead of the `user_id`, while maintaining the ability to retrieve customer information.
Option C: Encrypt the `user_id` field within the response body.
Option D: Remove the `user_id` field from the API response entirely.
Generating a unique random identifier (Option B) is a standard anonymization technique that effectively removes the direct link to the customer while still allowing for retrieval of their information. Encryption (Option C) doesn't address the re-identification risk; and simply removing the field (Option D) might break other functionalities.
Frequently Asked Questions
What does the "Anonymization Techniques" exercise practise?
Learn the IT-English vocabulary of data anonymisation: pseudonymisation, masking, aggregation, k-anonymity and re-identification.
How many questions are in this exercise?
This exercise has 22 questions, each multiple-choice with a full explanation shown after you answer.
What English level is this exercise for?
This exercise is tagged Intermediate. If the vocabulary feels difficult, browse the Data Privacy category page for an easier module to start with.
Is this exercise free to use?
Yes. Every exercise on CoderSlingo, including this one, is free with no account, sign-up, or paywall.
Do I get feedback if I answer incorrectly?
Yes — whichever option you choose, right or wrong, you'll immediately see an explanation clarifying the correct term and why the other options don't fit.
Can I retry this exercise?
Yes — once you finish all the questions, a "Try again" button on the results screen resets the exercise so you can practise as many times as you like.
Do I need an account to track my progress?
No account is required. Your progress bar and score for this session are tracked in the browser as you go, but nothing is saved once you leave the page.
Is "Anonymization Techniques" part of a larger series?
Yes — it's one exercise in the Data Privacy category on CoderSlingo. See the category page for the full list of related exercises on similar terminology.
Can I link directly to this exercise?
Yes — this exercise has its own permanent URL, so you can bookmark it or share the link directly with a colleague or study partner.
Where can I find more exercises like this one?
See the Data Privacy category page for related exercises, or browse the main Exercises hub for other IT English topics.