Full Data De-Identification

₹12500-37500 INR

Posted 1 minute ago

•

Ends in 6 days

₹12500-37500 INR

Paid on delivery

I need an experienced data-privacy professional to take every dataset we currently hold—internal Gmail logs, the code repository, CRM lead records, email engagement metrics, and performance-marketing creator stats—and strip out anything that could identify a person or our company before we pass the information to outside partners. The primary goal is to produce share-ready files that preserve analytical value while ensuring names, email addresses, phone numbers, company details, and any other personal or proprietary fields are irreversibly masked or removed. I will rely on you to spot additional risk factors, select the right techniques (tokenisation, hashing, k-anonymity, differential privacy, etc.), implement them in repeatable code (Python, SQL, or the tools you recommend), and validate the results. Deliverables • Clean, fully de-identified datasets covering all five data sources • Reproducible scripts or notebooks with clear inline comments • A short technical brief explaining the methodology, fields affected, and any residual risk assessment • Validation report confirming removal or masking of direct and indirect identifiers Acceptance will be based on successful test re-identification attempts returning no personal or company information and your documentation enabling us to rerun the process on future data drops.

Project ID: 40828178

About the project

8 proposals

Open for bidding

Remote project

Active 12 secs ago

Place your bid

Bid amount Email address

Benefits of bidding on Freelancer

Set your budget and timeframe

It's free to sign up and bid on jobs

8 freelancers are bidding on average ₹18,043 INR for this job

User Avatar
Flag of SPAIN

Hi, there "SECURE DATA DE-IDENTIFICATION WITH REPRODUCIBLE PRIVACY VALIDATION" — the main goal is to make all five datasets safe for external sharing while preserving their analytical value and preventing personal or company information from being exposed. I would audit each data source, identify direct and indirect identifiers, and apply appropriate techniques such as tokenisation, hashing, k-anonymity, and statistical anonymisation using Python and SQL. I would also develop reusable scripts, validate the output through re-identification tests, assess residual privacy risks, and document all transformations to ensure the process can be repeated reliably for future data drops. Which formats are your five datasets currently stored in, and do you have any preferred privacy standards such as GDPR or ISO 27701? Looking forward to working with you.

User Avatar
Flag of INDIA

Hello, This is a solid approach, you're not just scrubbing obvious PII but also watching for indirect identifiers that could re-identify someone when combined. The trickiest part will be balancing privacy with utility, especially with metrics like email engagement where patterns might still reveal too much. I'd start by auditing each dataset for direct and indirect identifiers, then apply masking or hashing where needed. For CRM leads and creator stats, I'd likely use k-anonymity to group small categories without losing trends. The validation would involve test queries trying to reconstruct original data from the cleaned versions, checking for leaks. Everything would be scripted in Python with Pandas and SQL, so you can rerun it easily. Did something similar last month for a document chatbot where we had to scrub user-uploaded files while keeping searchable content intact. How strict is the trade-off between data utility and privacy for the marketing stats, can some aggregation be lossy, or do you need exact numbers?

内容来自 Freelancer News,版权归原作者所有。

阅读原文 ↗

用 HomepageTab 新标签页关注 RSS,打开浏览器就能看到更新。

安装 HomepageTab