Thursday, July 14, 2022

Thinking About Sociothermodynamics


Twitter data reveals global communication network
Jul 2020, phys.org

I'd really like to read this study more in depth, but it's paywalled. So here's the methods: 
  • The scientists used the mentions mechanism in Twitter data to map the flow of information around the world. A mention in Twitter occurs when a user explicitly includes another @username in their tweet. This is a way to directly communicate with another user but is also a way to retransmit or retweet content.
  • The investigators examined Twitter data from December 2013 and divided the world into 8,000 cells, each approximately 100 kilometers wide. 
  • They measured modularity, a value that quantifies distance between communities on a network compared to a random arrangement. They also investigated a quantity known as betweenness centrality, which measures the number of shortest paths through each node.
  • 16 significant global communities. Three large communities exist in the Americas: an English-speaking region, Central and South American countries, and Brazil in its own group. Multiple communities exist in Europe, Asia and Africa.
  • Countries who had a common colonizer have a decreased preference of interaction (and suggests  hierarchical interactions with the colonizing country might inhibit interactions between former colonies).

via the MIT Media Lab and New England Complex Systems Institute: Leila Hedayatifar et al, Geographical fragmentation of the global network of Twitter communications, Chaos: An Interdisciplinary Journal of Nonlinear Science (2020). DOI: 10.1063/1.5143256


Geography can become a root cause for inequality when cities are built in a way that fragments social networks
Feb 2021, phys.org

We've known this with the Twitter mountain studies from years ago, but alas:
  • Urban planning directly influences the formation of social networks in a city 
  • Subsequently the socio-economic equality or inequality of its citizens
  • Social relations provide individuals with essential access to resources, information, economic opportunities and other forms of support
  • In towns with more evenly distributed social networks, the economic inequality tended to be much lower 
  • Dataset from Hungary, 2 million individuals from 500 towns via a social media platform used by 40% of the population 
  • "You cannot force people to interact if they don't want to", yet, towns and cities frequently make decisions about the built environment that will have effects on how their inhabitants can meet and interact

via Complexity Science Hub Vienna: G. Toth, J., R. Di Clemente, A. Jakobi, B. Sagvari, J. Kertesz, B. Lengyel, "Inequality is rising where social network segregation interacts with urban topology," Nature Communications (18 Feb 2021) DOI: 10.1038/s41467-021-21465-0


Mobility data reveals universal law of visitation in cities
May 2021, phys.org

The Sante Fe Institute is good at using scaling laws to determine the number of visitors to any location in a city. They use only these two variables -- how far visitors travel and how often they visit. 

"Imagine you are standing on a busy plaza, say in Boston, and you see people coming and going. This may look pretty random and chaotic, but the law shows that these movements are surprisingly structured and predictable. It basically tells you how many of these people are coming from 1, 2 or 10 kilometers away and how many are visiting once, twice or 10 times a month", says lead author Markus Schläpfer of ETH Zurich's Future Cities Laboratory. "And the best part is that this same regularity holds not only in Boston, but across cities worldwide."

Universally, they found that the number of visitors to any urban location scales as the inverse square of both travel distance from home and the visitation frequency. Like the gravitational pull of a large planet, an attractive city plaza with fine museums and famous shops draws relatively more visitors from more distant locations, though less frequently than those coming from nearby locations, their relative numbers being predictably determined by the inverse square law. A further surprising consequence of this new visitation law is that the same number of people visit the location whether they are coming from, say, 10km away 3 times a week, or from 3km away 10 times a week.

via the  Santa Fe Institute, MIT, and ETH Zürich: The universal visitation law of human mobility, Nature (2021). DOI: 10.1038/s41586-021-03480-9


Forget what you think you know about viral marketing, study suggests
Feb 2021, phys.org

Maybe advertising does work:

Even just a bit of advertising or other mass communication — "top-down" information that comes from outside the network — effectively equalizes the influence of everyone across the network [including the hub at the center, and against Malcolm Gladwell's network position priority described in Tipping Point.]

In instances where there is even a small amount of advertising — even when it is just a quarter of a percent as strong as word-of-mouth — there's virtually no difference between the influence of the person at the center of a network and those further out on the string.

"It takes only an incredibly weak amount of advertising to effectively neutralize the dominance of Mr. Popular." -UCLA professor of sociology Gabriel Rossman, who is also author of Climbing the Charts: What Radio Airplay Tells Us About the Diffusion of Innovation

via University of California, Los Angeles: Gabriel Rossman et al. Network hubs cease to be influential in the presence of low levels of advertising, Proceedings of the National Academy of Sciences (2021). DOI: 10.1073/pnas.2013391118


Communication technology, study of collective behavior must be 'crisis discipline'
Jun 2021, phys.org

Sociothermodynamics code red:
  • Researchers now say that the study of collective behavior must rise to a "crisis discipline," just like medicine, conservation and climate science have done.
  • Social media platforms are driven to maximize engagement and profitability, not to ensure sustainability or accurate information—and the vulnerability of these systems to misinformation and disinformation poses a dire threat to health, peace, global climate and more.
  • No one, not even the platform creators themselves, have much understanding of how their design decisions impact human collective behavior, the authors argue.
  • Good reminder: Historically collective behavior has best been understood as when animals or people exhibit coordinated action without an obvious leader. This includes how fish school to evade predators or when a crowd spontaneously breaks into applause or becomes silent.

via University of Washington: Joseph B. Bak-Coleman el al., "Stewardship of global collective behavior," PNAS (2021). www.pnas.org/cgi/doi/10.1073/pnas.2025764118


Cities like Paris may be optimal urban form for reducing greenhouse gas emissions
Aug 2021, phys.org

  • The built environment accounts for 39% of all greenhouse gas emissions in the U.S.
  • Low-rise, high-density environments are the optimal urban form when looking to reduce greenhouse gas emissions over their whole life cycle
  • Challenges current conventional understanding that tomorrow's cities must be densely packed and stretch upwards to address and curb greenhouse gas emissions
  • Development should focus on minimizing whole-life carbon of buildings, not just the emissions from their operations or their materials
  • Simulating 5,000 environments based on real-world data to establish their lifecycle greenhouse gas emissions
  • Results show that density is indeed needed for a growing urban population, but height isn't

via University of Colorado at Boulder with Edinburgh Napier University: Francesco Pomponi et al, Decoupling density from tallness in analysing the life cycle greenhouse gas emissions of cities, npj Urban Sustainability (2021). DOI: 10.1038/s42949-021-00034-w


Material efficiency holds great potential for climate neutrality
Aug 2021, phys.org

In the case of residential buildings, wood construction and reducing the living space per capita show the greatest savings potential. 

via University of Freiburg: Stefan Pauliuk et al, Global scenarios of resource and emission savings from material efficiency in residential buildings and cars, Nature Communications (2021). DOI: 10.1038/s41467-021-25300-4

Wooden Fractal on Metal FSK - Thom Fractal Forums - 2017

Scientists meld Twitter and satellite views to understand epic impact
Aug 2021, phys.org

  • Unique look at two enormous global projects that are part of China's Belt and Road Initiative, in part to understand if early concerns about environmental damage emerged, and if public sentiment turned critical
  • Examined satellite images of the vast areas the projects covered to see how forests, grassland, bodies of water and other land types changed
  • Nighttime light imagery as a way of gauging economic changes (more lights means more economic development)
  • Carefully chosen keywords and hashtags on Twitter to understand communication before, during and after the construction projects
  • "Our analysis shows that both projects have led to a substantial loss of natural land but gains in artificial land," said Yingjie Li, a Ph.D. candidate at MSU-CSIS who is the paper's lead author. "Despite this, we found that, overall, public sentiment toward the projects was largely positive and improved over time, which contradicts the prevalent pessimism by the Belt and Road Project critics."
  • People in more developed areas of Kenya and Pakistan were more positive about the sweeping changes than people in less-developed regions

Question though -- how did they make sure that the Twitter data was from people and not robots? I'm looking at you, human flesh engine. And I'm looking at any post on Reddit's World News that has the word China in the title, and how it has dozens of comments instead of thousands, because it's been smashed by robots and the moderators can't help but cull the real comments along with the fake. Case in point -- the comment section for a recent article about the closing of Macau due to covid numbers (circa July 2022), despite over 50 comments, had not one single instance of the words "China", or even "Macau" or "gambling" or "gaming", because the whole conversation on the sub had been manipulated by the Chinese human flesh engine to avoid any discourse about China itself. 

Looking at the paper itself, we see only reference to "0.4 million unique users" but no mention to whether they tried to discriminate between real users and fake ones. Also, there is no mention of fake accounts in the "Study limitations".

via Michigan State University's Center for Systems Integration and Sustainability: Yingjie Li et al, Synthesizing social and environmental sensing to monitor the impact of large-scale infrastructure development, Environmental Science & Policy (2021). DOI: 10.1016/j.envsci.2021.07.020


Socioeconomic networks and built environments of cities contribute to lower rates of psychological depression
Aug 2021, phys.org

Luís Bettencourt's work that shows how the infrastructure networks of cities — and how people move about such networks — can lead to rapid increases in social interaction and entails higher rates of innovation and wealth production.

One main finding is that, on average, people have more contacts across a greater variety of functions when they live in larger cities. Understanding that social isolation is a significant risk factor for depression, it made sense that more socialization through these varied networks could be protective against depression.

"Mental health and cognition are the basis for agency and behavior, and urban environments do change how people think and act. -Bettencourt

via University of Chicago: Andrew J. Stier et al, Evidence and theory for lower rates of depression in larger US urban areas, Proceedings of the National Academy of Sciences (2021). DOI: 10.1073/pnas.2022472118


Deep learning helps to predict traffic crashes before they happen
Oct 2021, phys.org

Typically, these types of risk maps are captured at much lower resolutions that hover around hundreds of meters, which means glossing over crucial details since the roads become blurred together. These maps, though, are 5x5 meter grid cells, and the higher resolution brings newfound clarity: The scientists found that a highway road, for example, has a higher risk than nearby residential roads, and ramps merging and exiting the highway have an even higher risk than other roads.

Previous attempts to predict crash risk have been largely "historical," as an area would only be considered high-risk if there was a previous nearby crash.

To evaluate the model, the scientists used crashes and data from 2017 and 2018, and tested its performance at predicting crashes in 2019 and 2020. Many locations were identified as high-risk, even though they had no recorded crashes, and also experienced crashes during the follow-up years.

Their data: density, speed, and direction of traffic, and satellite imagery that describes road structures, such as the number of lanes, whether there's a shoulder, or if there's a large number of pedestrians

via MIT's Computer Science and Artificial Intelligence Laboratory and and the Qatar Center for Artificial Intelligence: Inferring high-resolution traffic accident risk maps based on satellite imagery and GPS trajectories.


A gender dimension of energy: Modern cooking fuels connected to quicker demographic transition
Dec 2021, phys.org

Switching to modern cooking fuels like gas or to electricity can improve the well-being of women in the global South, and eventually be connected to falling birth rates. ... Because it not only improves health, it also relieves women of the need to have many children to do time-consuming housework like fetching firewood or cooking on open fires. This frees up time to seek information and education—and eventually helps women realize their reproductive rights. This is a direct line connecting the switch from modern energies to the demographic transition,"

via Potsdam Institute for Climate Impact Research: Helga Weisz, Fertility transition powered by women's access to electricity and modern cooking fuels, Nature Sustainability (2021). DOI: 10.1038/s41893-021-00830-3.


Physicists unify sociological theories that explain social stability
Feb 2022, phys.org

Homophily is everything. 

And they use the 100,000-person community of a game called Pardus for data. 

"Based on homophilic one-to-one interactions, the computer simulation produced a society that almost magically self-organized towards more stability."

I just love seeing the word "magically" appear in science articles. 

via Complexity Science Hub Vienna and the Santa Fe Institute: Tuan Minh Pham et al, Empirical social triad statistics can be explained with dyadic homophylic interactions, Proceedings of the National Academy of Sciences (2022). DOI: 10.1073/pnas.2121103119


Immigration patterns are reflected in Facebook data on popular foods and drinks
Feb 2022, phys.org

Fucking genius:
Researchers have developed a novel strategy for using Facebook data to measure cultural similarity between countries, revealing associations between immigration patterns and people's food and drink interests. ... The method employs data on the top 50 food and drink preferences for any given country as captured by the Facebook Advertising Platform. ... For instance, the top 50 foods and drinks from Mexico are more popular in the U.S. than the top 50 U.S. foods and drinks are in Mexico, reflecting a greater degree of immigration from Mexico to the U.S. than vice versa.
But also, there's cultural evolutionary science in them data, because it isn't just food and drinks that are moving disproportionately between the countries. The metabolism of the anthroposphere.

via Public Library of Science, Max Planck Institute for Demographic Research in Rostock, Germany: Vieira CC, Lohmann S, Zagheni E, Vaz de Melo POS, Benevenuto F, Ribeiro FN (2022) The interplay of migration and cultural similarity between countries: Evidence from Facebook data on food and drink interests. PLoS ONE 17(2): e0262947. doi.org/10.1371/journal.pone.0262947


Using graph neural networks to measure the spatial homogeneity of road networks
May 2022, phys.org

  • 11,790 urban road networks in 30 cities worldwide, including cities from either developed countries or developing countries
  • "Traditional road network metrics based on simple measures only provide a coarse characterization of urban road networks."
  • They specifically trained the graph neural networks to learn the structure patterns of the road networks in the observed regions, so that they could then predict the structure of the network in the hidden region.
  • (Intercity and intracity homogeneity across all the cities is what they're looking for.)
  • The graph neural networks developed by this team of researchers could ultimately be used in different countries worldwide to measure the similarities between cities, evaluate urban policies and summarize urban activities. 

"Beyond road networks, the homogeneity theories of street views, land use, and other infrastructure networks could also be established. There is much more opportunity to find more deep insights into complex urban systems."

via Purdue University and Peking University: Jiawei Xue et al, Quantifying the spatial homogeneity of urban road networks via graph neural networks, Nature Machine Intelligence (2022). DOI: 10.1038/s42256-022-00462-y


Post Script:
Brookings Institution's Center on Social Dynamics and Policy, which applies complex systems science to the study of social dynamics and their implications for policy, mainly through the use of computational modeling and simulation. 

Santa Fe Institute, which focuses on complexity theory.

Complexity Science Hub Vienna, which is the focal point of complexity science in Europe (and "free of bureaucratic constraints for open-minded visionaries who are brave enough to step out of mainstream science"!)

New England Complex Systems Institute, which is dedicated to advancing analytics and its application to the challenges of society, and the interaction of complex systems with the environment.


No comments:

Post a Comment