Search

The pitfalls of centralized social networks, such as Facebook and Twitter/X, have led to concerns about control, transparency, and accountability. Decentralized social networks have emerged as a result with the goal of empowering users. These decentralized approaches come with their own tradeoffs, and therefore multiple architectures exist. In this paper, we conduct the first large-scale analysis of Bluesky, a prominent decentralized microblogging platform. In contrast to alternative approaches (e.g. Mastodon), Bluesky decomposes and opens the key functions of the platform into subcomponents that can be provided by third party stakeholders. We collect a comprehensive dataset covering all the key elements of Bluesky, study user activity and assess the diversity of providers for each sub-components., Comment: To be presented at IMC'24

Large Language Models (LLMs) have demonstrated remarkable capabilities in comprehending and analyzing lengthy sequential inputs, owing to their extensive context windows that allow processing millions of tokens in a single forward pass. However, this paper uncovers a surprising limitation: LLMs fall short when handling long input sequences. We investigate this issue using three datasets and two tasks (sentiment analysis and news categorization) across various LLMs, including Claude 3, Gemini Pro, GPT 3.5 Turbo, Llama 3 Instruct, and Mistral Instruct models. To address this limitation, we propose and evaluate ad-hoc solutions that substantially enhance LLMs' performance on long input sequences by up to 50%, while reducing API cost and latency by up to 93% and 50%, respectively., Comment: 11 pages, 5 figures, 6 tables

How similar are politicians to those who vote for them? This is a critical question at the heart of democratic representation and particularly relevant at times when political dissatisfaction and populism are on the rise. To answer this question we compare the online discourse of elected politicians and their constituents. We collect a two and a half years (September 2020 - February 2023) constituency-level dataset for USA and UK that includes: (i) the Twitter timelines (5.6 Million tweets) of elected political representatives (595 UK Members of Parliament and 433 USA Representatives), (ii) the Nextdoor posts (21.8 Million posts) of the constituency (98.4% USA and 91.5% UK constituencies). We find that elected politicians tend to be equally similar to their constituents in terms of content and style regardless of whether a constituency elects a right or left-wing politician. The size of the electoral victory and the level of income of a constituency shows a nuanced picture. The narrower the electoral victory, the more similar the style and the more dissimilar the content is. The lower the income of a constituency, the more similar the content is. In terms of style, poorer constituencies tend to have a more similar sentiment and more dissimilar psychological text traits (i.e. measured with LIWC categories).

Out-of-context news is a common type of misinformation on online media platforms. This involves posting a caption, alongside a mismatched news image. Existing out-of-context news detection models only consider the scenario where pre-labeled data is available for each domain, failing to address the out-of-context news detection on unlabeled domains (e.g. news topics or agencies). In this work, we therefore focus on domain adaptive out-of-context news detection. In order to effectively adapt the detection model to unlabeled news topics or agencies, we propose ConDA-TTA (Contrastive Domain Adaptation with Test-Time Adaptation) which applies contrastive learning and maximum mean discrepancy (MMD) to learn domain-invariant features. In addition, we leverage test-time target domain statistics to further assist domain adaptation. Experimental results show that our approach outperforms baselines in most domain adaptation settings on two public datasets, by as much as 2.93% in F1 and 2.08% in accuracy.

The Universe is shaped as a web-like structure, formed by clusters, filaments, and walls that leave large volumes in between named voids. Galaxies in voids have been found to be of a later type, bluer, less massive, and to have a slower evolution than galaxies in denser environments (filaments and walls). However, the effect of the void environment on their stellar population properties is still unclear. We aim to address this question using 118 optical integral field unit datacubes from the Calar Alto Void Integral-field Treasury surveY (CAVITY), observed with the PMAS/PPaK spectrograph at the 3.5m telescope at the Calar Alto Observatory (Almer\'ia, Spain). We used the non-parametric full spectral fitting code STARLIGHT to estimate their stellar population properties: stellar mass, stellar mass surface density, age, star formation rate (SFR), and specific star formation rate (sSFR). We analysed the results through the global and spatially resolved properties. Then, we compared them with a control sample of galaxies in filaments and walls from the CALIFA survey, matched in stellar mass and morphological type. Key findings include void galaxies having a slightly higher half-light radius (HLR), lower stellar mass surface density, and younger ages across all morphological types, and slightly elevated SFR and sSFR (only significant enough for Sas). Many of these differences appear in the outer parts of spiral galaxies in voids (HLR > 1), which are younger and exhibit a higher sSFR, indicative of less evolved discs. This trend is also found for early-type spirals, suggesting a slower transition from star-forming to quiescent states in voids. Our analysis indicates that void galaxies, influenced by their surroundings, undergo a more gradual evolution, especially in their outer regions, with a more pronounced effect for low-mass galaxies., Comment: 24 pages, 22 figures, 5 tables. Accepted for publication in A&A. Updated references list

The recent development of decentralised and interoperable social networks (such as the "fediverse") creates new challenges for content moderators. This is because millions of posts generated on one server can easily "spread" to another, even if the recipient server has very different moderation policies. An obvious solution would be to leverage moderation tools to automatically tag (and filter) posts that contravene moderation policies, e.g. related to toxic speech. Recent work has exploited the conversational context of a post to improve this automatic tagging, e.g. using the replies to a post to help classify if it contains toxic speech. This has shown particular potential in environments with large training sets that contain complete conversations. This, however, creates challenges in a decentralised context, as a single conversation may be fragmented across multiple servers. Thus, each server only has a partial view of an entire conversation because conversations are often federated across servers in a non-synchronized fashion. To address this, we propose a decentralised conversation-aware content moderation approach suitable for the fediverse. Our approach employs a graph deep learning model (GraphNLI) trained locally on each server. The model exploits local data to train a model that combines post and conversational information captured through random walks to detect toxicity. We evaluate our approach with data from Pleroma, a major decentralised and interoperable micro-blogging network containing 2 million conversations. Our model effectively detects toxicity on larger instances, exclusively trained using their local post information (0.8837 macro-F1). Our approach has considerable scope to improve moderation in decentralised and interoperable social networks such as Pleroma or Mastodon., Comment: Accepted at International AAAI Conference on Web and Social Media (ICWSM) 2024. Please cite accordingly!

An important concept in organisational behaviour is how hierarchy affects the voice of individuals, whereby members of a given organisation exhibit differing power relations based on their hierarchical position. Although there have been prior studies of the relationship between hierarchy and voice, they tend to focus on more qualitative small-scale methods and do not account for structural aspects of the organisation. This paper develops large-scale computational techniques utilising temporal network analysis to measure the effect that organisational hierarchy has on communication patterns within an organisation, focusing on the structure of pairwise interactions between individuals. We focus on one major organisation as a case study - the Internet Engineering Task Force (IETF) - a major technical standards development organisation for the Internet. A particularly useful feature of the IETF is a transparent hierarchy, where participants take on explicit roles (e.g. Area Directors, Working Group Chairs). Its processes are also open, so we have visibility into the communication of people at different hierarchy levels over a long time period. We utilise a temporal network dataset of 989,911 email interactions among 23,741 participants to study how hierarchy impacts communication patterns. We show that the middle levels of the IETF are growing in terms of their dominance in communications. Higher levels consistently experience a higher proportion of incoming communication than lower levels, with higher levels initiating more communications too. We find that communication tends to flow "up" the hierarchy more than "down". Finally, we find that communication with higher-levels is associated with future communication more than for lower-levels, which we interpret as "facilitation". We conclude by discussing the implications this has on patterns within the wider IETF and for other organisations.

Statistical learning methods are widely utilized in tackling complex problems due to their flexibility, good predictive performance and its ability to capture complex relationships among variables. Additionally, recently developed automatic workflows have provided a standardized approach to implementing statistical learning methods across various applications. However these tools highlight a main drawbacks of statistical learning: its lack of interpretation in their results. In the past few years an important amount of research has been focused on methods for interpreting black box models. Having interpretable statistical learning methods is relevant to have a deeper understanding of the model. In problems were spatial information is relevant, combined interpretable methods with spatial data can help to get better understanding of the problem and interpretation of the results. This paper is focused in the individual conditional expectation (ICE-plot), a model agnostic methods for interpreting statistical learning models and combined them with spatial information. ICE-plot extension is proposed where spatial information is used as restriction to define Spatial ICE curves (SpICE). Spatial ICE curves are estimated using real data in the context of an economic problem concerning property valuation in Montevideo, Uruguay. Understanding the key factors that influence property valuation is essential for decision-making, and spatial data plays a relevant role in this regard.

The structure of magnetic fields in galaxies remains poorly constrained, despite the importance of magnetism in the evolution of galaxies. Radio synchrotron and far-infrared (FIR) polarization and polarimetric observations are the best methods to measure galactic scale properties of magnetic fields in galaxies beyond the Milky Way. We use synthetic polarimetric observations of a simulated galaxy to identify and quantify the regions, scales, and interstellar medium (ISM) phases probed at FIR and radio wavelengths. Our studied suite of magnetohydrodynamical cosmological zoom-in simulations features high-resolutions (10 pc full-cell size) and multiple magnetization models. Our synthetic observations have a striking resemblance to those of observed galaxies. We find that the total and polarized radio emission extends to approximately double the altitude above the galactic disk (half-intensity disk thickness of $h_\text{I radio} \sim h_\text{PI radio} = 0.23 \pm 0.03$ kpc) relative to the total FIR and polarized emission that are concentrated in the disk midplane ($h_\text{I FIR} \sim h_\text{PI FIR} = 0.11 \pm 0.01$ kpc). Radio emission traces magnetic fields at scales of $\gtrsim 300$ pc, whereas FIR emission probes magnetic fields at the smallest scales of our simulations. These scales are comparable to our spatial resolution and well below the spatial resolution ($<300$ pc) of existing FIR polarimetric measurements. Finally, we confirm that synchrotron emission traces a combination of the warm neutral and cold neutral gas phases, whereas FIR emission follows the densest gas in the cold neutral phase in the simulation. These results are independent of the ISM magnetic field strength. The complementarity we measure between radio and FIR wavelengths motivates future multiwavelength polarimetric observations to advance our knowledge of extragalactic magnetism., Comment: Accepted for publication in the ApJ. Replaced to match final version. 35 pages, 17 figures, 4 tables

Simulation and emulation are popular approaches for experimentation in Computer Networks. However, due to their respective inherent drawbacks, existing solutions cannot perform both fast and realistic control plane experiments. To close this gap, we introduce Horse. Horse is a hybrid solution with an emulated control plane, for realism, and simulated data plane, for speed. Our decoupling of the control and data plane allows us to speed up the experiments without sacrificing control plane realism.

On the 21st of February 2022, Russia recognised the Donetsk People's Republic and the Luhansk People's Republic, three days before launching an invasion of Ukraine. Since then, an active debate has taken place on social media, mixing organic discussions with coordinated information campaigns. The scale of this discourse, alongside the role that information warfare has played in the invasion, make it vital to better understand this ecosystem. We therefore present a study of pro-Ukrainian vs. pro-Russian discourse through the lens of Twitter. We do so from two perspectives: (i) the content that is shared; and (ii) the users who participate in the sharing. We first explore the scale and nature of conversations, including analysis of hashtags, toxicity and media sharing. We then study the users who drive this, highlighting a significant presence of new users and bots.

From health to education, income impacts a huge range of life choices. Earlier research has leveraged data from online social networks to study precisely this impact. In this paper, we ask the opposite question: do different levels of income result in different online behaviors? We demonstrate it does. We present the first large-scale study of Nextdoor, a popular location-based social network. We collect 2.6 Million posts from 64,283 neighborhoods in the United States and 3,325 neighborhoods in the United Kingdom, to examine whether online discourse reflects the income and income inequality of a neighborhood. We show that posts from neighborhoods with different incomes indeed differ, e.g. richer neighborhoods have a more positive sentiment and discuss crimes more, even though their actual crime rates are much lower. We then show that user-generated content can predict both income and inequality. We train multiple machine learning models and predict both income (R-squared=0.841) and inequality (R-squared=0.77)., Comment: The Paper is accepted for publication in ICWSM 2023

We present the analysis of the magnetic field ($B$-field) structure of galaxies measured with far-infrared (FIR) and radio (3 and 6 cm) polarimetric observations. We use the first data release of the Survey on extragALactic magnetiSm with SOFIA (SALSA) of 14 nearby ($<20$ Mpc) galaxies with resolved (5 arcsec-18 arcsec; $90$ pc--$1$ kpc) imaging polarimetric observations using HAWC+/SOFIA from $53$ to $214$ \um. We compute the magnetic pitch angle ($\Psi_{B}$) profiles as a function of the galactrocentric radius. We introduce a new magnetic alignment parameter ($\zeta$) to estimate the disordered-to-ordered $B$-field ratio in spiral $B$-fields. We find FIR and radio wavelengths to not generally trace the same $B$-field morphology in galaxies. The $\Psi_{B}$ profiles tend to be more ordered with galactocentric radius in radio ($\zeta_{\rm{6cm}} = 0.93\pm0.03$) than in FIR ($\zeta_{\rm{154\mu m}} = 0.84\pm0.14$). For spiral galaxies, FIR $B$-fields are $2-75$\% more turbulent than the radio $B$-fields. For starburst galaxies, we find that FIR polarization is a better tracer of the $B$-fields along the galactic outflows than radio polarization. Our results suggest that the $B$-fields associated with dense, dusty, turbulent star-forming regions, those traced at FIR, are less ordered than warmer, less-dense regions, those traced at radio, of the interstellar medium. The FIR $B$-fields seem to be more sensitive to the activity of the star-forming regions and the morphology of the molecular clouds within a vertical height of few hundred pc in the disk of spiral galaxies than the radio $B$-fields., Comment: 28 pages, 14 figures. Accepted for publication in The Astrophysical Journal

We study the influence of different activation functions in the output layer of deep neural network models for soft and hard label prediction in the learning with disagreement task. In this task, the goal is to quantify the amount of disagreement via predicting soft labels. To predict the soft labels, we use BERT-based preprocessors and encoders and vary the activation function used in the output layer, while keeping other parameters constant. The soft labels are then used for the hard label prediction. The activation functions considered are sigmoid as well as a step-function that is added to the model post-training and a sinusoidal activation function, which is introduced for the first time in this paper., Comment: Accepted in ACL 2023 SemEval Workshop as selected task paper

The acquisition of Twitter by Elon Musk has spurred controversy and uncertainty among Twitter users. The move raised as many praises as concerns, particularly regarding Musk's views on free speech. As a result, a large number of Twitter users have looked for alternatives to Twitter. Mastodon, a decentralized micro-blogging social network, has attracted the attention of many users and the general media. In this paper, we track and analyze the migration of 136,009 users from Twitter to Mastodon. Our analysis sheds light on the user-driven pressure towards centralization in a decentralized ecosystem and identifies the strong influence of the social network in platform migration. We also characterize the activity of migrated users on both Twitter and Mastodon.

As an alternative to Twitter and other centralized social networks, the Fediverse is growing in popularity. The recent, and polemical, takeover of Twitter by Elon Musk has exacerbated this trend. The Fediverse includes a growing number of decentralized social networks, such as Pleroma or Mastodon, that share the same subscription protocol (ActivityPub). Each of these decentralized social networks is composed of independent instances that are run by different administrators. Users, however, can interact with other users across the Fediverse regardless of the instance they are signed up to. The growing user base of the Fediverse creates key challenges for the administrators, who may experience a growing burden. In this paper, we explore how large that overhead is, and whether there are solutions to alleviate the burden. We study the overhead of moderation on the administrators. We observe a diversity of administrator strategies, with evidence that administrators on larger instances struggle to find sufficient resources. We then propose a tool, WatchGen, to semi-automate the process.

Mergers are thought to be a fundamental channel for galaxy growth, perturbing the gas dynamics and the magnetic fields (B-fields) in the interstellar medium (ISM). However, the mechanisms that amplify and dissipate B-fields during a merger remain unclear. We characterize the morphology of the ordered B-fields in the multi-phase ISM of the closest merger of two spiral galaxies, the Antennae galaxies. We compare the inferred B-fields using $154~\mu$m thermal dust and $11$ cm radio synchrotron emission polarimetric observations. We find that the $154~\mu$m B-fields are more ordered across the Antennae galaxies than the $11$ cm B-fields. The turbulent-to-ordered $154~\mu$m B-field increases at the galaxy cores and star-forming regions. The relic spiral arm has an ordered spiral $154~\mu$m B-field, while the $11$ cm B-field is radial. The $154~\mu$m B-field may be dominated by turbulent dynamos with high $^{12}$CO(1-0) velocity dispersion driven by star-forming regions, while the $11$ cm B-field is cospatial with high HI velocity dispersion driven by galaxy interaction. This result shows the dissociation between the warm gas mainly disturbed by the merger, and the dense gas still following the dynamics of the relic spiral arm. We find a $\sim8.9$ kpc scale ordered B-field connecting the two galaxies. The base of the tidal tail is cospatial with the HI and $^{12}$CO(1-0) emission and has compressed and/or sheared $154~\mu$m and $11$ cm B-fields driven by the merger. We suggest that amplify B-fields, with respect to the rest of the system and other spiral galaxies, may be supporting the gas flow between both galaxies and the tidal tail., Comment: 11 pages, 5 figures, Accepted for publication in ApJ Letters

Recent years have witnessed growing consolidation of web operations. For example, the majority of web traffic now originates from a few organizations, and even micro-websites often choose to host on large pre-existing cloud infrastructures. In response to this, the "Decentralized Web" attempts to distribute ownership and operation of web services more evenly. This paper describes the design and implementation of the largest and most widely used Decentralized Web platform - the InterPlanetary File System (IPFS) - an open-source, content-addressable peer-to-peer network that provides distributed data storage and delivery. IPFS has millions of daily content retrievals and already underpins dozens of third-party applications. This paper evaluates the performance of IPFS by introducing a set of measurement methodologies that allow us to uncover the characteristics of peers in the IPFS network. We reveal presence in more than 2700 Autonomous Systems and 152 countries, the majority of which operate outside large central cloud providers like Amazon or Azure. We further evaluate IPFS performance, showing that both publication and retrieval delays are acceptable for a wide range of use cases. Finally, we share our datasets, experiences and lessons learned., Comment: 14 pages, 11 figures

The "Decentralised Web" (DW) is an evolving concept, which encompasses technologies aimed at providing greater transparency and openness on the web. The DW relies on independent servers (aka instances) that mesh together in a peer-to-peer fashion to deliver a range of services (e.g. micro-blogs, image sharing, video streaming). However, toxic content moderation in this decentralised context is challenging. This is because there is no central entity that can define toxicity, nor a large central pool of data that can be used to build universal classifiers. It is therefore unsurprising that there have been several high-profile cases of the DW being misused to coordinate and disseminate harmful material. Using a dataset of 9.9M posts from 117K users on Pleroma (a popular DW microblogging service), we quantify the presence of toxic content. We find that toxic content is prevalent and spreads rapidly between instances. We show that automating per-instance content moderation is challenging due to the lack of sufficient training data available and the effort required in labelling. We therefore propose and evaluate ModPair, a model sharing system that effectively detects toxic content, gaining an average per-instance macro-F1 score 0.89.

Decentralising the Web is a desirable but challenging goal. One particular challenge is achieving decentralised content moderation in the face of various adversaries (e.g. trolls). To overcome this challenge, many Decentralised Web (DW) implementations rely on federation policies. Administrators use these policies to create rules that ban or modify content that matches specific rules. This, however, can have unintended consequences for many users. In this paper, we present the first study of federation policies on the DW, their in-the-wild usage, and their impact on users. We identify how these policies may negatively impact "innocent" users and outline possible solutions to avoid this problem in the future.

Current IP based networks support a wide range of delay-sensitive applications such as live video streaming of network gaming. Providing an adequate quality of experience to these applications is of paramount importance for a network provider. The offered services are often regulated by tight Service Level Agreements that needs to be continuously monitored. Since the first step to guarantee a metric is to measure it, delay measurement becomes a fundamental operation for a network provider. In many cases, the operator needs to measure the delay on all network links. We refer to the collection of all link delays as the Link Delay Vector (LDV). Typical solutions to collect the LDV impose a substantial overhead on the network. In this paper, we propose a solution to measure the LDV in real-time with a low-overhead approach. In particular, we inject some flows into the network and infer the LDV based on the delay of those flows. To this end, the monitoring flows and their paths should be selected minimizing the network monitoring overhead. In this respect, the challenging issue is to select a proper combination of flows such that by knowing their delay it is possible to solve a set of a linear equation and obtain a unique LDV. We first propose a mathematical formulation to select the optimal combination of flows, in form of ILP problem. Then we develop a heuristic algorithm to overcome the high computational complexity of existing ILP solvers. As a further step, we propose a meta-heuristic algorithm to solve the above-mentioned equations and infer the LDV. The challenging part of this step is the volatility of link delays. The proposed solution is evaluated over real-world emulated network topologies using the Mininet network emulator. Emulation results show the accuracy of the proposed solution with a negligible networking overhead in a real-time manner.

The worldwide spread of COVID-19 has prompted extensive online discussions, creating an `infodemic' on social media platforms such as WhatsApp and Twitter. However, the information shared on these platforms is prone to be unreliable and/or misleading. In this paper, we present the first analysis of COVID-19 discourse on public WhatsApp groups from Pakistan. Building on a large scale annotation of thousands of messages containing text and images, we identify the main categories of discussion. We focus on COVID-19 messages and understand the different types of images/text messages being propagated. By exploring user behavior related to COVID messages, we inspect how misinformation is spread. Finally, by quantifying the flow of information across WhatsApp and Twitter, we show how information spreads across platforms and how WhatsApp acts as a source for much of the information shared on Twitter.

DNS is a vital component for almost every networked application. Originally it was designed as an unencrypted protocol, making user security a concern. DNS-over-HTTPS (DoH) is the latest proposal to make name resolution more secure. In this paper we study the current DNS-over-HTTPS ecosystem, especially the cost of the additional security. We start by surveying the current DoH landscape by assessing standard compliance and supported features of public DoH servers. We then compare different transports for secure DNS, to highlight the improvements DoH makes over its predecessor, DNS-over-TLS (DoT). These improvements explain in part the significantly larger take-up of DoH in comparison to DoT. Finally, we quantify the overhead incurred by the additional layers of the DoH transport and their impact on web page load times. We find that these overheads only have limited impact on page load times, suggesting that it is possible to obtain the improved security of DoH with only marginal performance impact.

The study of animal movement is challenging because it is a process modulated by many factors acting at different spatial and temporal scales. Several models have been proposed which differ primarily in the temporal conceptualization, namely continuous and discrete time formulations. Naturally, animal movement occurs in continuous time but we tend to observe it at fixed time intervals. To account for the temporal mismatch between observations and movement decisions, we used a state-space model where movement decisions (steps and turns) are made in continuous time. The movement process is then observed at regular time intervals. As the likelihood function of this state-space model turned out to be complex to calculate yet simulating data is straightforward, we conduct inference using a few variations of Approximate Bayesian Computation (ABC). We explore the applicability of these methods as a function of the discrepancy between the temporal scale of the observations and that of the movement process in a simulation study. We demonstrate the application of this model to a real trajectory of a sheep that was reconstructed in high resolution using information from magnetometer and GPS devices. Our results suggest that accurate estimates can be obtained when the observations are less than 5 times the average time between changes in movement direction. The state-space model used here allowed us to connect the scales of the observations and movement decisions in an intuitive and easy to interpret way. Our findings underscore the idea that the time scale at which animal movement decisions are made needs to be considered when designing data collection protocols, and that sometimes high-frequency data may not be necessary to have good estimates of certain movement processes.

Under increasing scrutiny, many web companies now offer bespoke mechanisms allowing any third party to file complaints (e.g., requesting the de-listing of a URL from a search engine). While this self-regulation might be a valuable web governance tool, it places huge responsibility within the hands of these organisations that demands close examination. We present the first large-scale study of web complaints (over 1 billion URLs). We find a range of complainants, largely focused on copyright enforcement. Whereas the majority of organisations are occasional users of the complaint system, we find a number of bulk senders specialised in targeting specific types of domain. We identify a series of trends and patterns amongst both the domains and complainants. By inspecting the availability of the domains, we also observe that a sizeable portion go offline shortly after complaints are generated. This paper sheds critical light on how complaints are issued, who they pertain to and which domains go offline after complaints are issued., Comment: The Web Conference 2019

Over the past decade, IXPs have been playing a key role in enabling interdomain connectivity. Their traffic volumes have grown dramatically and their physical presence has spread throughout the world. While the relevance of IXPs is undeniable, their long-term contribution to the shaping of the current Internet is not fully understood yet. In this paper, we look into the impact on Internet routes of the intense IXP growth over the last decade. We observe that while in general IXPs only have a small effect in path shortening, very large networks do enjoy a clear IXP-enabled path reduction. We also observe a diversion of the routes, away from the central Tier-1 ASes supported by IXPs. Interestingly, we also find that whereas IXP membership has grown, large and central ASes have steadily moved away from public IXP peerings, whereas smaller ones have embraced them. Despite all this changes, we find though that a clear hierarchy remains, with a small group of highly central networks

In the paper we analyze 26 communities across the United States with the objective to understand what attaches people to their community and how this attachment differs among communities. How different are attached people from unattached? What attaches people to their community? How different are the communities? What are key drivers behind emotional attachment? To address these questions, graphical, supervised and unsupervised learning tools were used and information from the Census Bureau and the Knight Foundation were combined. Using the same pre-processed variables as Knight (2010) most likely will drive the results towards the same conclusions than the Knight foundation, so this paper does not use those variables.