TLDR Starting from human perception, this post systematically unpacks the physical and physiological basis of “color”: visible-light spectrum, the L/M/S cone cells in the retina, and the separation of luminance and chroma, so readers can see why three dimensions suffice to describe color. It then covers metamerism and the difference between additive and subtractive mixing, clarifying how displays (emitted light) and print (pigment) impose different constraints in practice, and how the continuous spectrum is geometricized into a color wheel to support intuitive picking and palette strategies. On color spaces, it contrasts perception-oriented CIE-family spaces with implementation-oriented RGB-family spaces and the problems each solves. Finally it introduces a perception-friendly workspace, HCT, and how it delivers more natural gradients, more consistent palettes, and better alignment with accessibility standards in front-end and visualization work.

TLDR This article provides a systematic introduction to the core concepts, methodologies, historical evolution, and emerging trends in data visualization. It begins by elucidating the essence of visualization—optimizing information transmission through visual encoding while leveraging the human eye's high-bandwidth parallel processing capabilities to achieve rapid data comprehension. The article then traces the evolution of visualization from the 17th century to present day, along with a microcosm of China's charting library development from ECharts to AntV to VCharts. At the application level, it delineates three major domains: scientific visualization (simulation), information visualization (communication), and visual analytics (analysis), with particular focus on frontier topics including narrative visualization, AI4VIS, VIS4AI, and immersive visualization. Finally, it points out that by empowering the design pipeline with AI and introducing well-designed narrative interactions, the industry can effectively address the dual challenges of designer scarcity and high cognitive barriers for users.

TLDR This article explores Leland Wilkinson's The Grammar of Graphics theory in depth, which reduces visualization design from chart-level to more fundamental graphic elements. The article first introduces the core concepts of the grammar of graphics, including its specifications (data, transform, scale, coordinate, element, guide) and the three steps of assembly and display. It then compares the implementations of ECharts and AntV G2 to illustrate the different design philosophies and their trade-offs between chart-type-based and grammar-of-graphics-based approaches. Finally, the article reflects on the current core challenge in the visualization field - design itself rather than tools - and emphasizes the importance of professional visualization design knowledge and critical thinking capabilities through specific cases like extreme value handling and class grade reports in complex scenarios.

TLDR This article explores the fundamental definition and practical application of data visualization. Through examining the value of tabular visualization, it establishes that visualization's core purpose is "optimizing information transmission through visual encoding." Using bill visualization as a case study, the article details the complete process from design concept to implementation, covering technical aspects such as bar chart ordering, animation design, and layout composition. The final solution presents a multi-layered visualization based on a polar coordinate system, effectively displaying multi-dimensional bill data through inner, middle, and outer rings, with continuous optimization through practical implementation.

TLDR This article systematically reviews how to use AI coding assistants like Cursor efficiently and correctly. Key takeaways include: multi-turn conversations significantly reduce LLM accuracy—always prefer providing a complete, one-shot description of your requirements; optimizing your project structure and code semantics with an "AI-friendly architecture" (such as self-documenting code, contract-based design, structured READMEs, DSL term mapping, etc.) can greatly enhance AI understanding and collaboration; make full use of Cursor's various inquiry modes (Agent, Ask, Manual, Custom, Background) and context management features, flexibly switching according to your project's needs; and select the right AI model based on task type and model characteristics. The overarching message: treat code and documentation structure from an engineering perspective, proactively provide high-quality context for the AI, and you'll unlock the full value of AI coding assistants.

TLDR The HCT color space is a next-generation color model developed by Google, balancing human visual perception and design needs, and is widely used for dynamic theme color generation. Its utility library supports multiple languages, can automatically generate frontend-friendly theme palettes, and allows custom distribution functions for localization and charting scenarios. HCT's perceptually uniform color gradients and theme generation mechanism provide designers and developers with more natural and comfortable color experiences, though there is still room for optimization in localization and special applications.

TLDR Splines are smooth curves defined by piecewise polynomials, widely used in data visualization. Different types of splines have distinct characteristics: Bezier curves are easy to use but lack local control at higher orders; Hermite curves are based on physically meaningful velocity vectors; Cardinal and Catmull-Rom curves pass directly through control points; B-Splines achieve C² continuity while maintaining local control capability. Continuity (C^n and G^n) affects the smoothness and visual appearance of curves. For data visualization, choosing appropriate spline curves can balance accuracy and reduce visual illusions.

TLDR
This article introduces various 2D collision detection methods: from basic Axis-Aligned Bounding Boxes (AABB) and circular collision, to Oriented Bounding Boxes (OBB) and the Welzl algorithm for minimum enclosing circles; from the Separating Axis Theorem (SAT) for convex polygons, to precise pixel-level collision detection; finally covering performance optimization through Hierarchical Bounding Boxes (HBB) and spatial partitioning methods like quadtrees. Different methods are suitable for different scenarios, providing solutions for collision detection from simple geometric shapes to complex irregular figures.

TLDR
There are two classic methods for fluid simulation: the Eulerian method and the Lagrangian method. This article will provide a detailed explanation of the principles and implementation of both approaches.

TLDR

This article provides a detailed guide on how to localize Odinblade's series of class MODs for DOS2 into Chinese. The main steps include:

1. Using LSLib Toolkit to unpack the MOD's PAK files
2. Locating and translating localization files in the MOD, including content in Public/Localization and Mods/Localization
3. Using The Divinity Engine2 editor for deeper localization work, handling text not included in standard localization files
4. Repackaging and publishing the localization MOD

Through this method, the author successfully provided complete Chinese localization support for Odinblade's series of class MODs, gaining approximately 3000 subscribers.

TLDR

This article explores a new interaction method for complex network visualizations that combines fisheye distortion with eye tracking technology. The approach allows users to naturally explore dense node-link diagrams by controlling the fisheye effect with their gaze, while keeping the mouse available for other interactions. This is particularly useful for organizational relationship visualizations where users need to switch between overview and detailed views of specific connections.

TLDR

This article explores the gaming industry's evolution and current state. It examines key metrics showing the industry's massive scale ($240 billion) and rapid growth, with 3.42 billion gamers worldwide and revenues of $187.7 billion in 2024. The article traces the industry's transformation from a simple three-part structure (developers, publishers, platforms) to today's vertically integrated ecosystems dominated by major players like Sony, Nintendo, and Microsoft. It analyzes how subscription models have become a dominant revenue stream, similar to Netflix's impact on entertainment. The piece concludes by discussing mobile gaming's rise due to accessibility and fragmented gaming experiences, while predicting virtual reality headsets as the next major platform evolution once they overcome current technical barriers.

TLDR This article explores how to enhance data visualization using ControlNet-Canny, inspired by historical Zoomorphic Maps. By leveraging AI tools like ComfyUI and ControlNet, we can infuse charts with thematic elements while preserving their data integrity, creating more engaging and contextually relevant visualizations. The experiment demonstrates a modern approach to an age-old cartographic technique of making data more expressive and meaningful through creative visual representation.

TLDR This article documents my experimental attempt at creating pixel art animations using AI tools. Through exploring different approaches like ComfyUI, I tried to generate character sprite sheets for a simple running animation. While the results are still preliminary and face several challenges, this experiment provides some insights into combining AI with pixel art creation. The article shares my learning process and the technical workflow I developed, hoping it might be helpful for others interested in similar experiments.

TLDR

The study investigates the involvement of automation in narrative visualization tools, summarizing six genres: annotated charts, infographics, timelines & storylines, data comics, scrollytelling & slideshow, and data videos. It categorizes four types of tools:design spaces, authoring tools, ML/AI-supported tools, and ML/AI-generator tools according to their intelligence and automation levels. By surveying 105 papers and tools, the study explores how automation can progressively engage in visualization design and narrative processes to aid users in creating narrative visualizations. It aims to provide an overview of current research and development, identifies key research problems in each category, and suggests opportunities for further investigation in this domain.

TLDR

This article provides a comprehensive overview of artificial intelligence (AI), covering its fundamental concepts, historical development, and key milestones. Starting with a basic definition of intelligence as a system that maps inputs to specific outputs, it explores important events like the 1956 Dartmouth Conference that established AI as an academic field, and the emergence of connectionism in 1958 which introduced neural network approaches. The article includes detailed explanations, diagrams, and mathematical concepts to help understand AI's theoretical foundations and practical implementations.

TLDR

This article explores how ECharts implements morphing transitions between different shapes. We'll examine the core implementation in the source code, breaking down the process into key steps including path conversion, structure alignment, and optimal rotation calculation. Through this analysis, we'll understand how ECharts achieves smooth shape transformations while maintaining visual continuity.

TLDR

This article explores how to extend ECharts' global roaming functionality. While ECharts provides roaming through dataZoom for Cartesian charts, relationship-type charts lack full global roaming support. We'll examine ECharts' current single-series roaming implementation and develop a solution to enable global roaming across all chart types.

TLDR

LIDA is a novel tool for generating grammar-agnostic visualizations and infographics. It addresses several key challenges in automatic visualization creation:

1. Understanding data semantics
2. Enumerating visualization goals
3. Generating visualization specifications

The system uses a multi-stage pipeline powered by large language models (LLMs) and image generation models (IGMs). It consists of 4 main modules:

1. SUMMARIZER: Converts data into compact natural language summaries
2. GOAL EXPLORER: Identifies potential visualization goals based on the data
3. VISGENERATOR: Handles visualization code generation, refinement and filtering
4. INFOGRAPHER: Creates data-driven stylized graphics using IGMs

LIDA provides both a Python API and an interactive user interface supporting direct manipulation and multilingual natural language for generating charts, infographics and data stories.

TLDR

This article contains my notes about font rendering.