The workshop focuses on enhancing climate resilience in Georgia through the Georgia Climate Services for Agriculture (GECSA). Funded by the Green Climate Fund and supported by the Swiss and Swedish governments and UNDP, this initiative aims to reduce communities' exposure to climate-driven disasters. By improving agrometeorological services, the program provides farmers with tailored climate and weather information, promoting climate-smart practices.  

Key players, including national agricultural agencies and farmers, collaborate to generate user-friendly climate data. The presentation will highlight the development of a weather observation network, featuring innovative 3D printing technology for weather stations. This initiative showcases a multidisciplinary approach to integrating data and technology, creating sustainable tools for agriculture in the face of climate change.

Large Language Models (LLMs) are driving innovation in AI, revolutionizing the way complex tasks are addressed across multiple sectors. This talk will explore the core principles behind the development of LLM-based intelligent systems, with a focus on best practices in Prompt Engineering, Retrieval-Augmented Generation (RAG), Fine-Tuning, and System Design. Additionally, it will cover comprehensive evaluation methodologies designed to ensure these systems achieve high performance and reliability, while also aligning with business objectives and ethical standards.

Gabriel Geiger, Investigative Reporter at Lighthouse Reports, will explore the innovative approach to scrutinizing secretive algorithms that affect welfare decisions across Europe. He will share how his team gained unprecedented access to an AI system designed to predict fraud among welfare recipients, tested it for bias, and co-published their findings with WIRED magazine. Learn how journalists and civil society can obtain meaningful access to opaque AI systems and craft compelling narratives from rigorous technical analyses that engage mainstream audiences.

You’ve been analyzing your organization’s financial data and have just discovered several valuable insights, which you decide to communicate to the management team in a series of charts. Instead of getting “oohs” and “aahs,” though, you get questions like, “What am I supposed to get from this?” and “Why are these charts so hard to read?” Or, worse, silence. Or, even worse, questions from the audience that indicate that they completely misinterpreted the charts.  In this eye-opening talk, globally recognized data visualization instructor and best-selling author Nick Desbarats will reveal the real reasons why so many charts flop with audiences — even charts from billion-dollar high-tech companies, universities, and major news media outlets — and how chart creators can learn to design charts that are easy to read, professional looking, and compelling.

Each year, states, international organizations, and research institutions around the world produce hundreds of documents of relevance to human rights and sustainable development. These files include tens of thousands of observations, analysis, comments, and recommendations, challenging the human capacity to digest comprehensively. The insights from all this information can have revealing implications for the understanding of social phenomena and the investigation of domestic, regional, and global trends to inform decision-making. By extracting meaningful patterns and information from large volumes of unstructured textual data, text mining can play a crucial role in enabling more effective strategies for addressing the intricate challenges faced by marginalized populations. Concretely, it will discuss the experiences and potential of using the text mining algorithm from the Danish Institute for Human Rights to derive insights from large volumes of text and inform issues on sustainable development, human rights and rightsholder groups.

As data professionals, we all want to harness the power of storytelling to create compelling data narratives. However, most data storytelling techniques are designed for PowerPoint presentations,  which poses a challenge for dashboards, where all the visuals are displayed at once, and the data as well as insights might be still unknown at the moment of report creation.  But there are ways how you can create a story-like experience in a dashboard. Martynas will remind us the fundamental principles of storytelling, and then he'll show us how to apply those principles to lay out our data story in a dashboard that's as close to the "story-reading" experience as possible. After this talk, you're going to approach dashboard design in a brand-new way!

Ana went from a highly structured academic career path to an unpredictable but very satisfying world of freelancing in the last few years. In this talk she will practice a form of nostalgia called advice and compile a few things she wishes she had known when starting out — things about expectations, legal and accounting issues, project hunting, mindset about skills and ownership of work. This talk is for anyone who has worked in data space for a few years and is looking for more freedom with their work as well as a higher risk/benefit awareness. Also, if you’re a time traveler, Ana is willing to give you her slides and spacetime coordinates.

Engage in discussions in informal settings with industry experts and data enthusiasts about data, tech, AI and ethics. Connect with fellow attendees for sharing ideas and experiences in your respective field.

This talk fills in the blank with buzzy words like 'Society', 'Ethics', 'Governance', 'Scale','Use Cases', 'The Tech Stack', and more to explore two perspectives on AI: (1) the top-down approach, which is critiqued as dull and (ultimately) anti-technology, and (2) a bottom-up approach, which is promoted as an intelligent, practical way to stay excited about this very fluid and diverse thing called AI. This talk pulls from personal experiences in industry, civil society, and government to show how to stay in love with AI after the honeymoon is over.