by jsendak | Aug 19, 2025 | Art
The Power of Art: Reflecting on History and Exploring Contemporary Themes
Throughout history, art has served as a powerful vehicle for expression, reflection, and communication. From the ancient cave paintings of Lascaux to the masterpieces of the Renaissance, art has played a crucial role in shaping our understanding of the world and ourselves. It has the ability to transcend language barriers, cultural divides, and time itself, offering a glimpse into the soul of humanity.
Today, in an increasingly digital and interconnected world, the role of art remains as important as ever. Artists continue to push boundaries, challenge norms, and provoke thought through their creations. With each brushstroke, sculpture, or photograph, they explore complex themes such as identity, politics, social justice, and the environment, inviting us to pause, reflect, and engage with the world around us.
SWANFALL ART at Mall Galleries
This month, SWANFALL ART returns to Mall Galleries in London with its annual exhibition, showcasing a diverse range of works from emerging and established artists. With a focus on contemporary themes and innovative techniques, the exhibition promises to be a thought-provoking experience for art enthusiasts and casual visitors alike.
As you wander through the galleries, take a moment to consider the power of art in shaping our past, present, and future. Allow yourself to be inspired, challenged, and moved by the beauty and complexity of the works on display. For in the hands of talented artists, art has the ability to transform, connect, and elevate the human experience.
SWANFALL ART returns to Mall Galleries, London this month with its annual exhibition.
Read the original article
by jsendak | Aug 19, 2025 | DS Articles
[This article was first published on
Open Analytics, and kindly contributed to
R-bloggers]. (You can report issue about the content on this page
here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don’t.
Cloud providers are no longer just offering traditional x86-based servers,
ARM-based servers are now becoming a serious alternative. And for good reason:
they’re often far more cost-effective and power-efficient than their x86
counterparts. For instance, we found a provider offering a server with 16
virtual cores and 32 GB of RAM for just €30 per month. To be realistic, most
providers charge between €100 and €200 per month for such a server. These
more expensive servers usually have less customers on a single server, improving
the real life performance. In this post, we’ll put one of these budget-friendly
ARM servers to the test. Of course, similar results can be achieved with an
Intel or AMD server. The goal is to set up a production-grade ShinyProxy
environment, designed to host a single public app, and see how many concurrent
users it can realistically handle. (And yes, the title probably gives a hint
already!).
Although this post focuses on a specific use case, the same principles also
apply when running multiple apps with authentication.
Server selection
The first step is to pick a cloud provider and order a server. Choose a
datacenter that’s located close to your end users to minimize latency and
improve performance. Most providers also allow you to resize your server later
by adding more CPU or RAM, so you can start small and scale up as demand grows.
Make sure the server comes with a public IP address, since there is no need for
a load balancer in this setup.
When it comes to the operating system, you can use any Linux distribution you
prefer. In this tutorial, we’ll use Ubuntu 24.04. Once your server is
provisioned, the next step is to connect via SSH. Usually your cloud provider
provides documentation that explains how to do this.
As a final prerequisite, you’ll need a domain name. This makes it easier to
access your server using a human-friendly address and is also necessary for
configuring TLS. You can use either a root domain (e.g. example.com) or a
subdomain (e.g. shinyproxy.example.com). In your domain’s DNS settings, create
an A record that points the domain or subdomain to your server’s public IP
address.
Installation
With the server up and running, the next step is to install the ShinyProxy
Operator. This tool takes care of the ShinyProxy installation process as well as
all the required side components, so you don’t have to manage them manually. For
detailed instructions, follow the
official documentation.
For the best setup, make sure to enable TLS and the monitoring stack. At this
point you should be able to login and start the demo app.
Inspecting the resource consumption of the app
With ShinyProxy deployed, you can deploy your app. Once again, the
documentation contains
everything you need to know. To keep things simple, you can build the Docker
image on the server. For example, clone your Git repository with all your code on the
server and build the image using Docker. This way, you don’t need a container
registry. After you added the app to the ShinyProxy configuration, you can start
it through ShinyProxy. In order to know how many concurrent
users are supported by the server, it’s a good idea to first have a look at the
resources consumed by the app. Start the app, perform some user actions and run
the following command:
sudo docker stats
The output will be similar too:
The last line in the screenshot shows the container running our Shiny app. It
uses about 125 MiB of RAM and very little CPU time. Keep in mind that we’re
working with a small demo app, CPU usage will increase if your app processes
more data or performs heavier computations. Most Shiny apps typically use less
than 500 MiB of RAM, but some can consume multiple gigabytes, especially when
loading large datasets. To keep our example realistic, we’ll assume each app
requires 500 MiB of RAM. The other system components (including ShinyProxy) also
consume some memory. In the screenshot, these consume about 1.5 GiB. ShinyProxy
generally requires up to 2 GiB of RAM. To avoid degrading server performance,
it’s a good practice to leave a few gigabytes free. For this setup, we’ll
reserve 4 GiB for ShinyProxy and the server itself, leaving 28 GiB available for
the Shiny containers. With each container using 0.5 GiB, this allows us to run
up to 56 containers concurrently.
Optimizing the configuration
Originally, ShinyProxy created a new container for each user. While this approach
is simple and improves security, it’s not ideal for hosting a public app with
many users. Fortunately, since quite some time, ShinyProxy has excellent
support for what we
call pre-initialization and container-sharing.
With this feature, a single container can serve multiple users simultaneously.
Since the containers are pre-initialized, loading times are minimal. However,
each Shiny container still has a limit on the number of users it can handle. For
simple dashboards, you can typically expect about 20 concurrent users per
container. In our example, we’ll assume 10 concurrent users per container. Given
our 56 containers, this setup allows the server to handle up to 560 concurrent
users. To stay on the safe side and avoid overloading the server, we’ll round
this down to 500 concurrent users.
With these numbers in mind, it’s time to adjust the configuration. Let’s start
with the app configuration.
proxy:
specs:
- id: my-app
container-image: openanalytics/shinyproxy-demo
seats-per-container: 10
max-total-instances: 500
minimum-seats-available: 500
The seats-per-container: 10 setting tells ShinyProxy that each container can
handle 10 users or “seats”. To prevent overloading the server, we limit the
total number of concurrent users to 500 with the max-total-instances: 500
setting. If more than 500 users try to access the app at the same time, they
will see a message indicating that not enough capacity is available. Additionally, the
minimum-seats-available: 500 setting ensures that ShinyProxy pre-creates 50
containers and keeps them running, so users can start the app immediately
without waiting for containers to initialize.
In this example, we estimated the maximum number of concurrent users and limited
the server to that value. Since we’re deploying only one app, it makes sense to
create all containers at startup. However, if you are hosting multiple apps on a
single server (which ShinyProxy fully supports), it’s better to start with just a
few containers per app. ShinyProxy automatically creates new containers as
users begin accessing the apps, ensuring resources are used efficiently.
While you are updating the configuration of ShinyProxy, add (or update) the
following settings as well:
proxy:
hide-navbar: true
landing-page: SingleApp
authentication: none
# ...
memory-limit: 2Gi
The first property hides the ShinyProxy navbar, which is often unnecessary when
hosting a single public app. By setting landing-page: SingleApp, users are
automatically redirected to the app when they visit the landing page. Since this
is a public app, we also disable authentication. Finally, we limit the memory
that ShinyProxy can use to ensure there is always enough RAM available for the
Shiny apps.
After updating the configuration, the operator creates a new ShinyProxy instance
with the updated settings. This process usually takes a few minutes. You might
see a “forbidden” page if you previously logged in using a username and
password. This happens only once when switching from authentication: simple to
authentication: none. To fix it, simply clear the ShinyProxy cookies in your
browser.
As soon as ShinyProxy starts, it creates 50 new containers. When you launch your
app, it loads almost instantly, thanks to the pre-initialized containers. We
tested this setup with 500 browser sessions, and the system handled them
effortlessly. Both ShinyProxy and the Shiny app responded quickly, and the server
load remained minimal. Of course, actual performance depends on the resource
consumption of your app, but ShinyProxy is designed to manage these variations
efficiently. The setup demonstrates that even a single ARM-based server can
easily manage hundreds of concurrent users, delivering robust, production-grade
performance.
The first panel in the screenshot shows how long users wait before being
assigned a seat, and the second panel shows the number of available seats. When
we launched 500 sessions within just two minutes, the available seats dropped
quickly. Even so, loading times stayed at only a few milliseconds, which makes
it clear that ShinyProxy remains highly responsive during a sudden surge of
hundreds of new sessions.
Once all apps are loaded, ShinyProxy’s CPU usage drops and remains low, as shown
by the docker stats output:
Monitoring
Although the monitoring stack was enabled when ShinyProxy was first deployed,
disabling authentication means that Grafana is no longer accessible through
ShinyProxy. A simple workaround is to deploy a second ShinyProxy instance that
still uses authentication. This private instance can also serve as a testing
environment for new development versions of your app before deploying them to
the public production server. To set this up, create a new configuration file in
the input directory, for example private.shinyproxy.yaml, and add the
following code:
server:
servlet:
context-path: /private
Be sure to update the proxy.realm-id, for example by setting it to private.
After deploying the server, you can access it at the same hostname as the main
server, but under the /private/ sub-path. For instance, if your server is
available at http://localhost/, the private instance is accessible at
http://localhost/private/, and Grafana at http://localhost/private/grafana/.
To view the logs and metrics of the public instance, you need to adjust the
namespace variable in the dashboards.
Conclusion
With the introduction of the ShinyProxy Operator for Docker, setting up a
production-grade ShinyProxy server has never been easier. This post demonstrates
that only minimal configuration is needed to handle a large number of concurrent
users. Expensive servers or clusters aren’t required, just a single server is
sufficient. If you need to support more users, simply scale up the server: in
most cases, doubling CPU and RAM roughly doubles the number of concurrent users.
Compared to the resource usage of the Shiny apps themselves, ShinyProxy consumes
very little CPU and RAM, so it rarely becomes a bottleneck even under heavy
load.
Don’t hesitate to send in questions or suggestions and have fun with ShinyProxy!
Continue reading: Affordable Shiny app hosting for 500 concurrent users
Long-Term Implications and Future Developments
The shift towards using ARM-based servers instead of traditional x86-based servers has key long-term implications. These servers have the potential to provide cost-effective and power-efficient solutions for users, which will inevitably drive competitiveness among cloud providers and ultimately lower costs. Furthermore, the adoption of these servers will open more opportunities for app developers, as hosting solutions such as ShinyProxy become more affordable and accessible.
Potential ARM Success in the Cloud Server Market
ARM-based servers have a promising future in the cloud server market as they provide a significant cost-effective and power-efficient alternative to traditional x86-based servers. This may encourage cloud service providers to invest more in ARM architecture, resulting in advanced and more efficient cloud services. The expected growth of ARM-based servers further points towards a potential drop in the price of server hosting, making it affordable for startups and small business owners. This affordability will also serve to drive competitiveness among providers, benefiting users and the market as a whole.
The Future of Application Hosting
As demonstrated through the deployment of ShinyProxy on an ARM-based server, platforms such as ShinyProxy have the potential to revolutionize the manner applications are hosted. With the possibility of limiting the maximum number of concurrent users and pre-creating containers at startup, ShinyProxy and similar tools will provide app developers with unprecedented control over their hosted applications. Furthermore, the efficient use of resources by such platforms means that even single servers can support high-traffic applications, and scalability becomes seamless and cost-effective.
Actionable Advice
Consider the Shift to ARM-based Servers
If you are a business owner considering an upgrade or installation of new servers, you should contemplate ARM-based servers. Their cost-effectiveness and power-efficiency will help lower your overheads and improve performance. This shift might also future-proof your organization for potential cloud service improvements centered around ARM servers.
Utilize Affordable Application Hosting Solutions
ShinyProxy’s ability to efficiently manage applications on a server makes it a favorable tool for developers. Its efficient management of resources means that businesses can host their applications without requiring expensive server or cloud solutions. If affordability and high-traffic capability are vital factors for your application hosting needs, consider deploying your apps through ShinyProxy.
Prepare for the Future
Given the potential growth and future impact of ARM-based servers, staying informed about this technology could be crucial for your enterprise. Proactively incorporating modern and future-proofed tools like ShinyProxy into your technology stack can place your business in a better position to adapt to future industry changes and trends.
Read the original article
by jsendak | Aug 19, 2025 | DS Articles
This article explains how to install and setup Couchbase and start easily storing data.
Analyzing the Key Points of Couchbase Installation and Setup
By examining the information provided in the previous piece regarding the installation and setup of Couchbase, we can uncover potential long-term implications and future advancements. This document will aim to provide additional insights and advice based on these possible future scenarios.
Extended Implication and Future Developments
Couchbase, as a NoSQL database, offers distributed architecture designed to perform at scale. This should instigate future advancements aimed at improving overall performance, scalability, flexibility, and agility. The following are possible future developments.
- Enhanced Performance: Couchbase is likely to introduce performance enhancements to support the growing number of businesses embracing big data analytics.
- Increased Scalability: As data volumes grow, Couchbase will have to improve its scalability features to accommodate this surge.
- Upgraded Security: With growing incidences of cybercrime, future versions of Couchbase are expected to provide advanced security features.
Potential Long-term Implications
“The continuous use and development of Couchbase potentially has several long-term implications.”
- A Shift towards NoSQL Databases: More businesses could transition to NoSQL databases like Couchbase given its advanced features.
- Emergence of New Services: There could be an increase in services centered around Couchbase, such as consulting, maintenance, and support.
- Increased Demand for Couchbase Skills: As more organizations implement Couchbase, there will be a greater demand for professionals with Couchbase skills.
Actioning Insights for Future Success
To capitalize on these implications and developments, here are some strategies organizations and individuals might consider to stay ahead.
- Up-skilling or Re-skilling: Organizations should invest in training their workforce on Couchbase while individuals should consider acquiring Couchbase skills due to its growing demand.
- Invest in Couchbase Managed Services: To minimize the complexity of managing and optimizing the database, businesses should consider investing in Couchbase managed services.
- Plan for scalability: Businesses should prepare their infrastructures to handle the increased data volumes and performance needs that will come with advancing Couchbase versions.
By staying attuned to the ongoing developments of Couchbase and strategically implementing the above insights, organizations and individuals can be better poised for success in the dynamic tech landscape.
Read the original article
by jsendak | Aug 19, 2025 | DS Articles
Discover how AI is transforming financial forecasting for SMBs and enterprises from various industries. From real-time insights to predictive scenario planning, AI tools are helping companies improve accuracy, reduce risk, and adapt faster to market shifts.
The Transformation of Financial Forecasting through AI
Artificial Intelligence (AI) is significantly impacting various sectors, including the financial markets. For small and medium-sized businesses (SMBs) to large enterprises, AI plays an essential role in revolutionizing financial forecasting. AI tools’ abilities range from providing real-time insights to carrying out predictive scenario planning, key factors in improving accuracy, minimizing risks, and facilitating rapid adaptation to changes in the market. The importance of AI in finance is increasingly becoming apparent, but its long-term implications and potential future advancements are worth examining.
Future Developments and Long-term Implications
As AI’s capacities continue to evolve, it’s critical to explore these tools’ capability to alleviate several financial forecasting challenges. Companies utilizing AI can anticipate a transformative shift in their approach towards financial planning and forecasting.
“AI tools are the future of financial market analysis and decision making. They have the potential to make complex procedures simple, accurate, and time-effective” -Unknown Technology Expert
SMBs and enterprises investing in AI technologies may find significant returns on investment (ROI) over time. AI can unlock greater analytical power for financial forecasting, making it a vital tool in strategic business planning. Implementation of AI in financial forecasting signifies a move towards data-driven decision making, which can encompass everything from day-to-day operations to long-term strategic planning.
Advancements in AI for Financial Forecasting
- Real-time Insights: AI can process and analyze vast amounts of data almost instantly, providing real-time insights into market trends and customer behaviors. This speed and accuracy can help businesses stay ahead of their competitors.
- Predictive Scenario Planning: AI has the ability to simulate multiple financial scenarios based on historical data. This opens up opportunities for companies to prepare for various market situations, facilitating better business agility and resilience.
- Risk Management: AI can identify patterns and trends in data that human analysts may overlook. This data analysis can enable more accurate risk assessment, thereby reducing potential losses and improving overall financial security.
Actionable Advice for SMBs and Enterprises
For SMBs and enterprises combing over the horizon for a competitive edge in financial forecasting, seamless integration, and utilization of AI tools should be a top priority. Here are some steps for companies to move towards leveraging AI for financial forecasting:
- Invest in AI Tools: Companies should start by identifying and investing in suitable AI tools for their financial forecasting needs. This requires an assessment of their current data analysis capabilities and how AI might enhance them.
- Training and Skill Development: To efficiently use AI tools, businesses must ensure their staff possess the necessary skills to interpret AI outputs. This involves relevant training and development programs.
- Continual Evaluation: Once implemented, the effectiveness of these AI tools should be regularly checked. Through periodic evaluation, companies can gain insights into how well the AI tool matches the business’s changing needs.
By harnessing AI’s power, SMBs and large-scale enterprises can profoundly transform their financial forecasting, enhancing precision, reducing risks, and enabling prompt responses to market changes.
Read the original article
by jsendak | Aug 19, 2025 | Computer Science
arXiv:2508.12020v1 Announce Type: new
Abstract: The Audio-to-3D-Gesture (A2G) task has enormous potential for various applications in virtual reality and computer graphics, etc. However, current evaluation metrics, such as Fr’echet Gesture Distance or Beat Constancy, fail at reflecting the human preference of the generated 3D gestures. To cope with this problem, exploring human preference and an objective quality assessment metric for AI-generated 3D human gestures is becoming increasingly significant. In this paper, we introduce the Ges-QA dataset, which includes 1,400 samples with multidimensional scores for gesture quality and audio-gesture consistency. Moreover, we collect binary classification labels to determine whether the generated gestures match the emotions of the audio. Equipped with our Ges-QA dataset, we propose a multi-modal transformer-based neural network with 3 branches for video, audio and 3D skeleton modalities, which can score A2G contents in multiple dimensions. Comparative experimental results and ablation studies demonstrate that Ges-QAer yields state-of-the-art performance on our dataset.
Expert Commentary: Exploring Human Preference and Quality Assessment for AI-generated 3D Human Gestures
The Audio-to-3D-Gesture (A2G) task holds significant potential for various applications in virtual reality, computer graphics, and beyond. However, current evaluation metrics like Fre’chet Gesture Distance and Beat Constancy may not accurately capture human preference for generated 3D gestures. This gap highlights the need to delve deeper into understanding human perception and developing objective quality assessment metrics for AI-generated 3D human gestures.
Multi-disciplinary in nature, this research bridges the fields of multimedia information systems, animations, artificial reality, augmented reality, and virtual realities. By introducing the Ges-QA dataset, which comprises 1,400 samples with multidimensional scores for gesture quality and audio-gesture consistency, the authors have laid a solid foundation for further exploration in this domain.
The inclusion of binary classification labels to determine emotional matching between generated gestures and audio adds another layer of complexity to the task. This dataset enables the development of a multi-modal transformer-based neural network with separate branches for video, audio, and 3D skeleton modalities. This approach allows for scoring A2G contents across multiple dimensions, providing a more comprehensive assessment of gesture quality.
The comparative experimental results and ablation studies presented in the paper showcase the effectiveness of the proposed Ges-QAer model, demonstrating state-of-the-art performance on the dataset. This research not only contributes to advancing the field of AI-generated 3D human gestures but also underscores the importance of incorporating human preference into evaluation metrics for such tasks.
Read the original article
