Our phones, wearable devices, the websites we browse, even data generated by hospitals, banks, and government entities add to the byte count. PriceWaterhouseCoopers estimates that human beings will create at least 44 zettabytes of data before 2021, enough to fill up 44 billion one terabyte iPad Pros.

Our day-to-day interactions with simple data are fairly straightforward. Phone pictures, for example, are edited on apps, shared through social media accounts, and stored locally on devices or uploaded to a cloud service. Where it gets trickier is when you want to do a bit more with a spreadsheet or several pages of data.

Billion dollar corporations like Google and Facebook that deal with massive amounts of data, or big data, continually invest in complex systems and infrastructure to streamline how data is processed and analyzed.

Big data and building novel, large-scale database systems have always been Azza Abouzied’s primary focus. The assistant professor of computer science at NYU Abu Dhabi received the Very Large Data Bases’ Test of Time Award for her paper that explores building a hybrid data analysis system that merges database systems technology with then nascent scalable data processing frameworks. But she also has a burning desire to make data more accessible to everyone; she wants analyzing or extracting data to be easy for those who don’t want to invest in expensive, difficult tools or learn complex programming languages.

“I felt that I've done all these things to help companies that have big data, and actually care about speed and performance when analyzing data, but what about other people? What about regular data scientists, regular data journalists who want to be able to specify how to analyze data or to visualize it in an easy way?”

These questions have led her to develop various tools that simplify the process of data extraction, analysis, and visualization.

“Last year we published Qetch, a tool which helps users search time series data such as stock prices, heart rhythms, or any variable that fluctuates with time by simply sketching the pattern they wish to find,” Abouzied says. 

Her team also built SEER in collaboration with IBM, which helps people extract data from text documents by simply highlighting snippets of data. It is now in the process of being rolled out in their commercial text processing program called SystemT. SEER, Abouzied explains, can help journalists search for crime patterns in FBI press releases by simply highlighting statements of interest such as “battery declined by 10 percent” or “homicide doubled”.

Abouzied is currently looking at how she can develop tools for the healthcare sector where patient data can be mined and analyzed to help doctors make fast, informed decisions. It’s not going to replace what doctors do, she says, but will instead “support physicians by recommending the next best course of action whether it is further testing or treatment, or both.”

Data Security

Systems and databases are becoming increasingly intertwined and on the flip side of this staggering amount of data is the challenge of keeping it secure. Data security has been in the spotlight for a few years now, even more so since political consulting firm Cambridge Analytica harvested millions of Facebook users’ data without their knowledge. But it has always been relevant, argues Christina Pöpper, who teaches courses on digital privacy and computer security. The assistant professor of computer science thinks it’s crucial to stay vigilant.

“We (researchers) always have to wear two hats —the attacker hat and the defender hat. And we always have to switch sides. Sometimes our research is purely on the attack side. At other times, we work on making systems more secure, so we are on the defence side,” Pöpper explains.

Her research on GPS spoofing attacks, for instance, looks at how aerial vehicles relying on satellite navigation or air traffic control systems can be compromised, and how they could be safeguarded using crowdsourcing. Her team’s latest work on mobile network security investigates loopholes in the LTE specification even though it’s considered a more secure standard than its predecessors.

“For a few years, our research has focused a lot on LTE security, the latest widely deployed mobile telephony standard. It was designed to close a number of security problems from previous generations like GSM and 3G. Despite solid counter measures and security protocols, we could reveal that there are still a number of security issues and you figure that out piece by piece,” Pöpper continues.

Even as technology advances at an unprecedented pace, so do digital attacks. Security is increasingly becoming even more important, Pöpper says, because “there is no solution that covers everything.” 

She jokes that she'll always have a job since there’s no guarantee of technology being completely secure. Updates should be eyed with trepidation, which is why Pöpper is always on the lookout for gaps in the security puzzle.

“You can't secure complex systems 100 percent, so there's always some risk involved. That’s just something you have to accept in a certain way but it should not prevent us from striving to make systems more secure. Otherwise, the attackers would take over.”