Edvancer's Knowledge Hub
Data Security Measures for Data Scientists
Data breaches have become a regular occurrence. It is almost every day that there is a breach. You would have heard of the Equifax credit data breach, which is one of the latest in a series of stories about major organizations’ data being compromised and exposed. The list is endless; Target’s customer credit card database, Anthem’s health insurance records, and why, even the Federal Office of Personnel Management’s background check forms has not been spared! This is only the tip of the iceberg; critical servers and databases all over the world are compromised every single day. The frequency of the data breaches is only increasing and is only likely to worsen with time.
Well, you have enough reason to be alarmed! Security of data is now no longer the exclusive domain specialty of database administrators and network engineers. In today’s scenario, just about everyone who creates, manages, analyzes, or even just has access to data is a potential point of failure in an organization’s data security plan. Thus, in case you handle sensitive data, (data which you would refuse to share freely with any random stranger on the Internet) then it’s your responsibility to ensure that you have undertaken adequate security measures to protect that data.
The syllabus of formal training programs/grad schools for data scientists cover the data security aspects scantily, if at all, and thus, this is unfamiliar territory for most data scientists.
However, that does not mean that we do not pay data security, the attention it rightly deserves, more so, since, when even a minor negligence in that grey area can undo all the good work that you have otherwise done. Now, where does one even begin, in case one wishes to take good care of one’s data?
I will now dwell upon a few of the data security best practices that every good data scientist should follow:-
Share this on
Follow us on
- Take only what you need. The good old cardinal principle of sharing data strictly only on a “need to know basis” happens to be the first rule of data security. Data not held, cannot be lost, in the first place. It’s only when a clear need arises, that justifies the risk, that one should collect sensitive data. Even in that case, only get the absolutely minimum data that’s essential to complete the task. Data scientists are known to be tempted to hoard much more data than needed, out of the anxiety that they might require it later. This redundant accumulation of sensitive data can trigger what would be a minor cyber-security incident into a major disaster: so as a rule, just don’t do it.
- Understand the data you have, and don’t keep data you don’t need anymore. Apply the same principles from Rule #1 above to your existing data, assuming you already have some data. It would make good sense to document the data you have by keeping a regular inventory of the data you have on hand, regularly analyze the sensitivity of each dataset, permanently erase data you don’t need, and take proactive steps to reduce the risks inherent in data held by you - for instance, by removing or altering unstructured text fields, which can hide potentially sensitive data like phone numbers and names. It would also be prudent to think, not only about your own interests, but also about others too, when you have data about other people, by putting yourself in their shoes.
- Data should always be encrypted, to the extent possible. Encrypting data is not the one stop solution; however, it’s a low-cost method of adding an extra layer of protection in case a hard drive or network connection gets compromised. Other than in certain applications which demand extremely high performance, encryption does not impede performance grossly, in today’s scenario. Sensitive data should therefore, be encrypted by default. The conflict of having a trade-off between performance and encryption is now redundant; there are plenty of high-performance applications and services with built-in encryption available now, for example, this is a standard feature in Microsoft’s Azure SQL Database.
Manu Jeevan
Manu Jeevan is a self-taught data scientist and loves to explain data science concepts in simple terms. You can connect with him on LinkedIn, or email him at manu@bigdataexaminer.com.
Latest posts by Manu Jeevan (see all)
- Python IDEs for Data Science: Top 5 - January 19, 2019
- The 5 exciting machine learning, data science and big data trends for 2019 - January 19, 2019
- A/B Testing Made Simple – Part 2 - October 30, 2018
Follow us on
Free Data Science & AI Starter Course
