{"id":20048505,"url":"https://github.com/lefteris-souflas/census-privacy-analysis","last_synced_at":"2026-04-11T22:42:55.370Z","repository":{"id":229863230,"uuid":"777849189","full_name":"Lefteris-Souflas/Census-Privacy-Analysis","owner":"Lefteris-Souflas","description":"Exploring US Census microdata, tackling privacy issues, and anonymization. Exercise A delves into quasi-identifiers, anonymization methods, identification risks, and differential privacy. Exercise B involves data loading, k-anonymity, histograms, adding noise for privacy, computing private averages, and analyzing privacy parameter impacts.","archived":false,"fork":false,"pushed_at":"2024-04-17T20:59:25.000Z","size":4090,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-01-12T20:33:36.443Z","etag":null,"topics":["data-anonymization","differential-privacy","encryption","gaussian-mechanism","gaussian-noise","gdpr","jupyter-notebook","k-anonymity","pseudonymization","python3","quasi-identifiers"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Lefteris-Souflas.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null}},"created_at":"2024-03-26T16:05:42.000Z","updated_at":"2024-04-17T21:01:39.000Z","dependencies_parsed_at":"2024-03-31T04:00:13.186Z","dependency_job_id":null,"html_url":"https://github.com/Lefteris-Souflas/Census-Privacy-Analysis","commit_stats":null,"previous_names":["codeninjatech/census-privacy-analysis","lefteris-souflas/census-privacy-analysis"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Lefteris-Souflas%2FCensus-Privacy-Analysis","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Lefteris-Souflas%2FCensus-Privacy-Analysis/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Lefteris-Souflas%2FCensus-Privacy-Analysis/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Lefteris-Souflas%2FCensus-Privacy-Analysis/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Lefteris-Souflas","download_url":"https://codeload.github.com/Lefteris-Souflas/Census-Privacy-Analysis/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":241478108,"owners_count":19969229,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["data-anonymization","differential-privacy","encryption","gaussian-mechanism","gaussian-noise","gdpr","jupyter-notebook","k-anonymity","pseudonymization","python3","quasi-identifiers"],"created_at":"2024-11-13T11:44:23.482Z","updated_at":"2026-04-11T22:42:50.323Z","avatar_url":"https://github.com/Lefteris-Souflas.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Census Privacy Analysis\n\nAssignment 2 for the Business \u0026 Privacy Issues in Data Analysis Course of AUEB's MSc in Business Analytics\n\n## Description\n\nThe US Census Bureau provides microdata to be used in research and applications. These data were deemed a threat to privacy, and in 2020 they were anonymized for the first time. In this exercise, you must study the last non-anonymized census and discuss the privacy threats.\nAn example of the microdata of the 2010 census can be found [here](https://www.census.gov/data/datasets/2010/dec/stateside-pums.html). For a full description of the dataset, you can read the [technical documentation](https://www2.census.gov/programs-surveys/decennial/2010/technical-documentation/complete-tech-docs/us-pums/pumsus.pdf).\n\nFor the exercise, please use the Delaware data available [here](https://www2.census.gov/census_2010/12-Stateside_PUMS/Delaware/).\n\n### How to use the dataset\n\n1. Download and decompress the zip file.\n2. The extracted directory contains the `de.2010.pums.01.txt` file.\n3. The `de.2010.pums.01.txt` file contains rows corresponding to:\n   - Person records rows start with ‘P’\n   - Household records rows that start with ‘H’\n4. For the next steps ignore the Household records.\n5. For the Person records, please read the brief description for the comprised columns and the possible values for each column that can be found [here](https://www2.census.gov/census_2010/12-Stateside_PUMS/2010%20PUMS%20Record%20Layout.xlsx). We suggest using software like Excel or Numbers (manually) to guide the Person data or write a script to parse the file into a table.\n\n## Exercise A\n\nAfter examining the data table (columns \u0026 values), answer the following questions:\n1. Which attributes can act as **quasi-identifiers** and why?\n2. Which of the following properties holds for the data?\n   - They are anonymized\n   - They are pseudonymized\n   - They are encrypted\n- [x] Explain the key differences between the three approaches with respect to **GDPR**.\n3. Explain **how a person can be identified**.\n4. Define **differential privacy** and explain the importance of the **privacy parameter e**.\n\n## Exercise B\n\nLoad the dataset into a Python notebook (we suggest Jupyter) and display the first few rows to understand the data.\n1. Use the [Amnesia anonymization tool](https://amnesia.openaire.eu/) to apply **k-anonymity** to the dataset. Comment on the resulting dataset.\n2. Plot the distribution of numeric features in the dataset using histograms.\n3. Apply a **random noise mechanism** to some of the numeric columns using the Gaussian mechanism. The noise should be added to the original values in a way that preserves differential privacy.\n4. Calculate the differentially private averages for the individuals using the noisy data.\n5. Plot the distribution of numeric features after the noise addition. Try different values of the e parameter. Comment on the effect of the differential privacy on the results.\n\n- [x] For steps 2-5, you can use the **IBM** [differential privacy library](https://github.com/IBM/differential-privacy-library) and [notebooks](https://github.com/IBM/differential-privacy-library/tree/main/notebooks).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flefteris-souflas%2Fcensus-privacy-analysis","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Flefteris-souflas%2Fcensus-privacy-analysis","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flefteris-souflas%2Fcensus-privacy-analysis/lists"}