How has data science evolved to improve society?

by Jenny Darmody

AI thought leader Aruna Pattam discusses the wide-ranging applications of data science in society and the importance of diversity for the future of the industry.

Data science draws on a large array of skills and expertise within the tech sphere. It uses statistics and mathematics, analytical modelling, machine learning and AI to analyse data and extract insights.

Accenture’s Shane O’Neill said those who pursue a career in data science can potentially work in any industry that uses data, which nowadays is pretty much every industry.

But while we know that data science is important, you might not be aware of just how many areas of society it touches and how it can be used to solve a broad range of real-world challenges.

To gain a little more insight into the evolution of data science and its current uses, turned to Aruna Pattam. She is a global AI thought leader who currently works as the head of AI and data science at Indian multinational IT company HCL Technologies for the Asia Pacific and the Middle East region.

‘As a data scientist, I often am the only woman in the room’

Pattam has spent more than 20 years delivering platforms and decision support systems using analytics, artificial intelligence and machine learning.

“My career started with SAS where I learned how analytics could help solve business problems and create new opportunities,” she said.

“The more I learnt about what this field had to offer, the more passionate I became about its potential, which led me on a path in exploring analytics to transform financial services industry.”

The evolution of data science

While data science as a concept has been around for a long time, Pattam said the term itself only really came to be around 2008.

“In the early days, it was statisticians who developed algorithms to do analytics using probability models or machine learning algorithms.”

From there, Pattam saw analytics used for operational purposes such as business intelligence and data warehousing. “Here analytics was focussed on the speed of the algorithm, as opposed to its underlying statistical foundation,” she said.

“Then I have seen engineers working in transactional systems developing analytic models to guide the development of new features. Here analytics was concerned with data management issues, as well as modelling and early implementations of machine learning algorithms.”

After that came the full term of data science, which incorporates a mix of all that came before it but with an emphasis on modelling and machine learning techniques.

Data science in the real world

Many people might be familiar with how data science is used for recommendation engines. They may also associate AI with facial recognition. But there are countless real-world applications that can help change industries and even society for the better.

For example in healthcare, data science allows large amounts of clinical data to be processed to identify a set of treatments for a particular patient along with likely side effects. A doctor can then use that information for personalised patient treatment.

“In chemical and drug discovery, [it can help] sort through and compare various properties of millions of potential small molecules, synthesise, test and optimise in lab experiments before selecting the eventual drug candidate for clinical trials,” Pattam said.

“In the energy sector, patterns identified by data science are helping to predict demand, improve performance, reduce costs, prevent system failures thus achieve greater efficiencies.”

Data science can also help to develop early warning systems in financial risk management, health settings and agriculture. It can even help predict the effects of different climate crisis mitigation or pandemic management strategies and highlight those most promising.

Pattam said data science can also help “identify mental health concerns at both a population and individual scale and enable, for example, forum moderators to identify individuals in need of rapid intervention”.

Data science can be used to monitor and adjust operations in areas such as clean energy, logistics and communications, track and communicate health information to the public, and create smart cities that make more efficient use of public services, better manage traffic and reduce climate impacts.

However, while the benefits of data science across all of these areas are powerful, the industry is not without its challenges, especially when combined with AI and machine learning.

Challenges in the data world

Bias in AI is a huge topic and one that needs to be widely discussed as it could lead to serious consequences. There have been several examples in recent years highlighting ethical problems with AI, including an MIT image library to train AI that contained racist and misogynistic terms and the controversial credit score system in China.

Pattam said one of the most important ways bias in AI can be addressed is to do so “at an early stage of the development process” by including ethical considerations from the beginning.

“Everyone has a role to play in making the use of AI ethical, unbiased and beneficial for everyone,” she added.

‘We need to ensure the use of data does not raise privacy issues’

“Industry should be aware of their responsibilities as they develop and implement new technologies. The public sector has a role as influencer as well as an enabler that can support unbiased implementation of AI systems through regulation and policy.

“Academia needs to develop industry-independent knowledge about the potential for bias in AI systems and share that with the public.”

As well as concerns around bias, there is also a challenge with data privacy. “We need to ensure the use of data does not raise privacy issues,” said Pattam.

“The way to address this is a whole-of-society response, where companies and government play their part, but individuals also take responsibility for managing access to their information.”

Diversity in data science

Diversity is a key issue for all tech disciplines and, while the number of women in the field continues to grow, Pattam said women are still a minority in the data science industry.

“As a data scientist, I often am the only woman in the room,” she said.

“Some of the reasons for this underrepresentation are that it is a male-dominated field, sexism in STEM fields, bias against mothers, women lack confidence, a lack of female role models and a lack of visibility making it difficult for women to reach to new opportunities. Women also face an uphill battle in terms of promotion and recognition within organisations.”

Pattam said women starting off in the field of data science should develop connections, join a community and seek out opportunities even if these seem impossible.

“Try to have a mentor, someone who is going to help guide you through your career and give you advice on how to get where you want to be.”

She added that diversity in every respect, not just in gender, is one of the biggest things she would like to see change in the data science space.

“I hope that more women get involved in data science because they have different perspectives that are often valuable. I also hope to see more people of colour get into the field because there are big gaps here too. The world is not just white and male, so I think it’s important for everyone to play a role in changing the face of data science,” she said.

“I also hope to see more young people get involved. Kids are the future and I think they should be given opportunities to explore, learn and discover new things. I would like to see data science become so openly available that it is included in every school curriculum.”