My mother worked the counters at the State Bank of India when she was my age. She counted bills, served customers, and moved between branches before returning to the mess at home. Her life was demanding, but it was legible. Mine, as a freelancer working from my own desk, is harder to decipher: a tangle of logins, deadlines, and dependencies that shift faster than I can track them.

This is not just a personal observation. Life is now lived within interconnected networks, among accumulated systems upon systems.

Complexity is the defining condition of our time. All our lives are marked by events in social networks, financial and public infrastructure, health, and commodity supply chains. Part of the complexity is interconnected crises—all interacting, all changing, none behaving in ways that traditional science is fully equipped to explain.

“At a time which some see as a breakdown of order and many anxieties abound, the embarrassment of complexity is widely felt. It results from the gap between feeling overwhelmed by the sheer scale and scope of problems confronting us, and the pressure to pretend that we can manage them,” writes Science, Technology, and Society Studies scholar Helga Nowotny in the book 43 Visions of Complexity.

Within interconnected networks that adapt as they go, when one piece shifts, the consequences ripple outward in ways that defy simple cause-and-effect thinking. The reductionist approach in science, which gave us discoveries in microbiology and particle physics, is no longer sufficient to understand our complex world. We need new tools we can apply, not at the micro level of individual actors, like cells or particles, nor at the macro level of the whole system, but at the meso level—between the two. We badly need to understand the workings of the networks that we live within.

Tackling real-world networks in an abstract way is a job for physics and mathematics. A growing number of polymaths, physicists, mathematicians and interdisciplinary data scientists are answering the call to study the science of complex adaptive networks. These networks may be biological, social, and even philosophical. Thankfully, there is a developing field called complexity science, and it’s coming of age.

Recently, as an EMBO Maria Leptin Science Journalism fellow, I plunged into this new-ish science at the Complexity Science Hub in Vienna, Austria. All summer of 2025, I asked 25-odd scientists: What is complexity science? And how do you do it?

During the many interviews at the Hub, I met scientists studying how diseases spread, how polarisation in society occurs, what chess-playing machines reveal about the human mind, how we can prevent habitat loss to protect biodiversity, what an efficient transport system looks like, trends in AI adoption, and weak links in global supply chains. Attempts to answer these complex questions were built on the foundational models from physics and math. Each researcher observed a real-world network in an abstract way.

Such an approach is incredibly useful. Work by complexity scientists at the Hub in Vienna has informed public decision-making on multiple occasions. In one project, researchers at the Hub worked with the Austrian Federal Railways to find a way to minimise delays on the train network. In another example, in 2020, a model co-developed and operated at the Hub was used by the Austrian government to forecast how many ICU beds needed to be reserved for COVID-19 patients as the pandemic raged on, week by week. These numbers, in turn, informed the length and stringency of lockdowns in Austria, in a crosstalk between the government and scientists enshrined in Austria’s COVID law. By the end of my residency, the promising scientific work I encountered had truly renewed my hope in evidence-based policy.

Stefan Thurner, the Director of the 10-year-old Complexity Science Hub, is a pioneer in the field, Stefan has a knack for explaining why the world needs complexity science to study social systems. “In the many complex systems we study, fundamentally, there are some shared dynamics that have to do with nothing else but how networks are connected. Before a big transition, the networks prepare a little bit,” he said in an interview, wriggling his connected hands. “You don’t notice that they are doing that, especially if you don’t look. When you look, you can see little rearrangements here and there as the system prepares, and then you have a BANG. Rapid transitions follow. And rapid transitions in social systems are always associated with crisis and catastrophe,” he said.

This view pinpoints the main reason we need such a science. Complexity scientists can help examine the precursors of catastrophe and suggest ways to possibly alter course.

“As a community, we should not give the impression that we are able to tell the future,” Stefan added. “Wherever possible, we can improve the situation by asking: What will happen when this or that changes? What happens when we change the temperature or social pressure, or say, the density in a network? What will be the consequence of it, or what will be, say, the critical temperature of this system?” said the physicist by training.

A nation of contradictions

As an Indian science writer, I felt a strong urge to see such a scientific approach to understanding and problem-solving back home. With a billion-plus people navigating growing inequality, caste, language, geography, and rapid digitalisation all at once, India is one of the most complex societies on earth. Its monsoons, power grids, railways, and public health networks are all complex systems we don’t yet fully understand. After returning to home ground, I continued my interviews—this time with India’s burgeoning complexity science community.

Sarika Jalan, a complexity scientist at IIT Indore, is a rising star from the Indian complexity community. She organised the 2025 NetSciX, an international school and conference of the International Network Society, and recently delivered a keynote at the parent conference NetSci in Boston. With some others, she is setting up the India chapter of the International Complexity Science Society. Sarika is among Indian researchers who have already made significant contributions to the theoretical side of complex systems theory and are eager to develop models using real-world data. In an online interview, she told me, “​India provides many different avenues in the study of complex systems. Whether we can actually utilise this kind of research for the country’s benefit is another question.”

Our problems of inequality, ecosystem collapse, disaster management and other such issues will not yield to siloed expertise. These are problems of interconnection. Sarika offers an example: “Suppose there is a threat of floods in Uttarakhand. It may not be possible to solve the problem of evacuations and infrastructure development locally. If you view it through the lens of complexity science, you can take a broader view of the interactions at play. You take all elements into account, including storms, land use, and population. We have various tools in the study of complex adaptive science, like non-linear dynamics and our agent-based models, that we can apply to support a strategy for minimal loss in the case of a flood occurring.”

Complexity science is especially good at understanding how elements shift in their interactions and which kind of interactions can lead to crises. In Sarika’s example, it is very useful to know the conditions under which the system floods or collapses. And from her scientific perspective, understanding the conditions that will lead to major shifts requires systems-level thinking. She is not alone in this sentiment.

“Deep mathematics translated to real engineering”

At IIT Madras, combustion engineer-turned-climate scientist R.I. Sujith has been busy understanding and even predicting the Indian monsoons using the Indian government’s Indian Monsoon Data Assimilation and Analysis (IMDAA) datasets, among others. Sujith is a senior researcher with several discoveries and patents under his belt. “But I am a baby in complexity science,” he said. He runs the 2023-launched Centre of Excellence for Studying Critical Transitions in Complex Systems, one of the at least two dedicated groups at IIT Madras that work on complexity science.

Sujith’s view of complexity science is overwhelmingly positive. “It is deep mathematics translated to real engineering,” he stated. Before his foray into complexity, Sujith was an engineering researcher who made several discoveries about a phenomenon known as combustion instability.

(Here’s Sujith explaining combustion instability as a complex system, and another video showing NASA’s learning curve with the phenomenon.)

He metamorphosed into a complexity researcher when he realised that the systems he was studying inside combustors in gas turbines and rocket engines were complex systems. After collaborations with ISRO and working on General Electric engines, he is dedicating his “second innings” to applying what he learned about complex systems to the Indian monsoon.

Recently, Sujith collaborated with a group of monsoon scientists whose work “yielded good results on predicting the seasonal average of Indian Summer Monsoon Rainfall 18 months in advance,” he says. He worked on the physical explanation behind this successful long-lead prediction using synchronisation theory, a part of complexity theory, which examines how different systems adjust their rhythms to operate together.

His use of complexity theory to understand the Indian monsoon is only getting started. “We are able to monitor and declare the onset of monsoon at every location in the country. We are working on a prediction scheme to predict it, possibly a couple of weeks in advance.”

Another model he worked on predicts the seasonal average for Northeast Monsoon in Tamil Nadu with a 10-month lead.

His excitement about the vast potential of complexity science was reflected in a 2016 paper on early detection of critical transitions in complex systems, in which he and the other authors were quoted in a Nature article, suggesting: “their observations can be applied to spot similar tipping points in other domains, such as finance.” Sujith was pleasantly surprised when the paper was cited by Dominic Cummins, then the chief advisor to British Prime Minister Boris Johnson, in an open call for scientists to serve as policy experts for the UK government.

Data woes in India

“We have a huge scope in India, but one of the obstacles in India is the difficulty in getting good-quality data,” Sarika remarked, interrupting my fever dream of complexity-backed decision-making in India.

She recollected the time when she tried to gather data to model the effects of storms on India’s power grid. “I require very basic data on the latitude and longitude of transformers,” she told me. She contacted several agencies that hold this data, but got no response. Eventually, she found a smaller online database and decided to work with it. “Then the problem was that there was no latitude and longitude in the data, only names of places,” she remembered. This meant her PhD students spent a month manually determining the latitudes and longitudes of those places. “The scholars in my lab had to spend their time cleaning up data rather than developing their critical ideas on complexity science,” Sarika added.

Indian complexity scientists like Sarika speak wistfully of their colleagues in Europe, North America, South Korea, New Zealand, and Singapore, where public agencies enable easy access to real-world data and dedicated institutes have been established to study complexity science.

The challenge of getting one’s hands on interoperable data (clear data that can be used for various purposes) to apply complex system theories in India begins with establishing a data infrastructure for scientific use. There have been efforts, but these don’t meet the scale needed to nurture a thriving complexity science community.

Consider the case of Jawaharlal Nehru University-affiliated health and agricultural economics researcher Anirban Chakravarty, who has been part of efforts to build such data exchange frameworks. Since his PhD days, he has been interested in inequality in India. He has been studying social disparities in India as a complex system with interacting and adapting influences of health, agriculture and economics. Outputs of his work include apps that help farmers make informed decisions about price volatility and market indicators.

He sums up his research premise in this way: “If I am a government agency, how do I ensure that there is less inequality and the distribution of wealth is uniform. We work with historical data that we collect from different agencies, and then we use machine learning techniques to do predictive analysis.”

A huge amount of data is required for such a task. And that has been a big challenge. Teaming up with scientists from the Institute of Mathematical Sciences in Chennai, the TERI School of Advanced Studies in Delhi, and others, Anirban recently applied for a project grant under the ANRF Mission AI for Science and Engineering to build a repository of databases on health and social indicators from different agencies like the National Sample Survey Office, National Family Health Survey (NFHS) and others. “We want to collate all the available data, clean it, and make it available to develop foundational models for what we are calling quantitative health economics,” he said.

If they receive the grant, the group hopes to prepare datasets for use by various stakeholders and policymaking agencies. Anirban suggested that discussions are ongoing in the community on how to facilitate data sharing and improve coordination among agencies.

“Data has been a challenge in India, whether it is health data or economic data,” he said. “There is a lack of data of different granularities…meaning that often, only country-level data is of good enough quality. It becomes difficult for us to compete with other advanced regions like Europe because we lack proper documentation. For economic indicators or social governance indicators, we do not have data at the state level, and it’s even worse, district-wise. If you want to do some statistical analysis or machine learning modelling, you have to train your data on the past or historical data, and that’s when we face the challenge that we don’t have very long time series,” he added.

A time series is one data type that can do wonders for some complexity scientists. Time series analysis helps them examine the conditions under which an event occurred in a network, and assess dominoes that fell as a result of that event. Niraj Kushwaha, a researcher from Boisar, Maharashtra, was completing his PhD work at the Hub when I met him. He had spent the last four years developing a statistical model to see how armed conflicts on the African continent are connected. “When you zoom out and look at the sites of armed conflicts on a map, there is a definite pattern. The dataset contains only a time series of armed conflict events and the locations where they occurred. And, hence, it is free from any ethno-political biases,” he told me.

For his research, Niraj took the map of Africa and broke it down into random blocks. He then encoded in his model a time series dataset of armed conflict events in the region that he got from an international, independent database. As he explained, I saw coloured dots appear on a map on his laptop. “These are in no way sporadic events, and heuristic approaches to understanding armed conflict might not be useful for managing or preventing them,” he found. Heuristic approaches are those that rely on experience, intuition, or practical rules of thumb to solve problems or make decisions, rather than following a fixed, systematic method.

His results show that conflicts follow the same pattern as avalanches: a single event can set off an immediate cascade of similar events, each one feeding the next “Conflict avalanches,” he calls them. Niraj agreed that one could, in principle, apply his model to any region of the world.

Is returning to India on the cards for him? I asked Niraj. “That would be a dream! But not any time soon,” he said. He added that our country needs systematic quantitative studies and a thriving community to support such work in solving our societal and economic problems.

Understand, then predict

In 2021, a team of researchers from IIT Kanpur and IIT Hyderabad came up with the Susceptible, Undetected though infected, Tested positive, and Removed Analysis (SUTRA) model of the then-raging COVID pandemic. “Natural immunity provides significantly better protection against infection than the currently available vaccines,” the model, which was endorsed by the Indian government, concluded.

Soon after the model was publicised, other researchers pointed out its flaws. Eventually, the model wrongly predicted that the pandemic and its impact on India would soon end. Researchers criticised the model for endorsing the government’s attempts to prematurely celebrate its COVID measures that, in fact, proved to be inadequate, resulting in the deaths of many Indians.

The failure of the SUTRA model serves as a cautionary tale. Several complexity scientists I spoke to warned against relying on complexity science for predictions. What it offers is arguably more useful: the ability to ask ‘what if’. This ability can help us see, before a system breaks, the quiet rearrangements that signal an incoming crisis. Using this information, researchers can promote resilience in our systems. Simply put, complexity science can equip good planning.

Complexity science is not a crystal ball, cautions Stefan Thurner. “By definition, it seeks to understand systems that one cannot predict. That implies that complexity science could not predict anything. And I think that’s right, since complexity science wants to understand systems… not predict them, but to manage them,” he said.

Being inherently interdisciplinary, complexity science is well-suited to our modern times. New research questions are increasingly adopting AI tools to clean and reconstruct datasets. Equipped with new tools and rich in theoreticians of complex adaptive systems, India’s Complexity Science community is poised to undertake fruitful scientific endeavours.

The world is not going to get simpler. And a country of India’s scale and ambition cannot afford to govern complexity with 20th-century instruments. As one of the most complex societies, India’s stakes in getting this right are enormous.

Aashima Dogra is a science writer; she co-founded the feminist science media portal thelifeofscience.com and is the co-author of the recent book Lab Hopping, which investigates the realities behind the gender gap in Indian STEM.

This story was supported and published by the independent science media portal thelifeofscience.com.

This story was supported and published by the independent science media portal thelifeofscience.com. It is part of the Labhoping Summer Science Trilogy 2026

Also Read | How science academies can enhance scientific temper in India

Also Read | How a photodetector will diagnose urinary infections faster