Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[IEEE Access 2022] A Comprehensive State-of-the-Art Survey on Data Visualization Tools: Research Developments, Challenges and Future Domain Specific Visualization Framework #51

Open
uhhyunjoo opened this issue Nov 24, 2022 · 8 comments

Comments

@uhhyunjoo
Copy link
Owner

@uhhyunjoo
Copy link
Owner Author

Abstract

  • Data visualization 은 meaningful 한 data insights 를 interactive 하고 effective 한 방식으로 demostration 하기 위한 powerful 한 skill 이다.
  • 이 서베이에서는, data visualization 의 various scopes, aspects, theories 를 identify, classify, investigate 하기 위해 2017-2022의 70 논문을 모았다.
  • 또한 health, IoT, business dashboard, urban traffic management, smart budiling, environmental data visualization 등과 같은 다양한 domains and fields 의 powerful applications of data visualization 을 조사했다.
  • 그러나, basic sota best interative visualization techniques, web-based tools, platforms 에 대한 comprehensive study 를 없다고 결론냈다.
  • 따라서, visualization 의 performance 를 향상시키기 위해, theoretical, analytical, statistical models and techniques 를 철저히 조사했다.
  • current primary and domain specific future chanlleges 또한 리뷰 됐다.
  • related future research directions and opportunies 또한 추천 됐다.

@uhhyunjoo
Copy link
Owner Author

Introduction

  • Effective and interactive data visualization : 모든 formats 에 통틀어서 중요하다.
    • abstract (raw) data 를 physical (actionable) data 로 convert 한다.
    • ex. shape, color, position, length, height, width
    • persuasive stories 를 clear, logical, smart, plausible 한 방식으로 present 하기 위함이다.
  • 이때, large numbers of data 는 analyzed promptly 된다. 이는 visualization 을 efficient 하고 interactive 하게 만들기 위함이다.
  • data 의 rows and columns 는 청중들에게 설득적인 스토리를 어필하기에 부적절하다.
  • visualization 을 draw 하는 aim : clear, smart, persuasive 한 방식으로, decision maker 가 결론을 빠르게 만들게 돕기 위함이다.
  • 비전문적인 고객들이, 그래프와 차트를 만들고 이에 따라 중요한 결정을 만들게 하게 할 수도 있다.

@uhhyunjoo
Copy link
Owner Author

uhhyunjoo commented Nov 25, 2022

3) Decision Making in Visualization

image

image

  • Data visualization 은 차트, 그래프, 맵 뿐만 아니라, insights 를 visualizae 하기 위해서도 쓰인다.
  • large datasets 의 hidden info 를 다루기 위한 중요한 method 이다.
  • 시각화는 연구자, 과학자, 산업가가 데이터를 이해하고, 예상되는 위험을 식별하여 가치 있는 결과를 통해 사전 예방적으로 완화하는 데 도움이 된다.
  • Scope 3 은 빅데이터 분석, 데이터의 효과적인 데이터 표시, 데이터 시각화, 에 대한 decision making techniques 와 approaches 를 향상시키기 위한 top 연구자와 데이터 과학자의 여러 기술과 모델을 highlight 한다.
  • [41] 은 visualization 의 new domain 에 대해 다뤘다. uncertainty and decision making for interactive visualization.
    • actual work : to interpret the datasets, to exploit the insight to assess effective decision making
    • visualization 의 uncertainty 를 limit 하는 것을 돕기도 했다.
  • [42] 는 visualization 에서 decision making 을 support 하는 supporting tool 을 제안했고, interactive visualization for users 를 implement 했다.
    • domain experts 를 support 하기 위해, visualizaion 을 evaluate 하기 위한 Design guideline 도 개발됐다.
  • [43] 은 quantitative tasks data 의 visualization 에 대해 automatic decision making 을 하기 위해 various methods 에 worked on 했다.
    • Automatic decision making 은 big data problems 에서 an evolvingi approach 이다. 그리고 quantitative data insights 를 monitor 하고 interpret 하는 걸 돕는다.
  • [45] 는 interactive data and visualization 에 대한 a decision-making model 을 제안했다.
    • 해당 모델은, better data visualizaion 을 위해, interactive visualizaion 에 대한 meaningful insights 를 share 하고, maximum features 에 대해 work on 하는 걸 support 할 수 있다.

@uhhyunjoo
Copy link
Owner Author

7) Applications of Visualization

image

image

@uhhyunjoo
Copy link
Owner Author

image

@uhhyunjoo
Copy link
Owner Author

II. Preliminaries of data visualization

A. STATE-OF-THE-ART VISUALIZATION TECHNIQUES

우선, 본 논문에서는 data visualization techniques 를 7개의 그룹으로 나눴다.

  • data distribution
  • data correlation
  • data ranking
  • data evolution
  • data maps
  • data flow

image

@uhhyunjoo
Copy link
Owner Author

III. EFFECTIVE TECHNIQUES FOR ENHANCING THE PERFORMANCE OF DATA VISUALIZATION

A. APPROXIMATE VISUAL ANALYTICS

image

  • hight dimensional info 에 대한 visual study 와 analysis 는 challenging 한 task 이다.
  • Direct visual analysis : scatterplot 이나 parallel coordiante 과 같은 few metrics 에서 works well gㅏㄴ다.
  • 그러나, high dimensional dataset 에 대해서는 not effective 하다.
  • Indirect visual analysis : high dimensionality challenges 에 대해 work and provide better performance 가능 하다.
  • Aggregation queries : also crucial class for the sequence of columns values
  • The main issues : to manage selective arbitary predicates, to offer through error guarantees without keeping the huge samples size.
  • [97] : the effective way to provide speedy answers to aggregation queries, [96] 은 measured biased sampling scheme 를 제안했다.
    • 또한 a solution for random samples aggregation 에 또한 제안했다.
  • [98] : to inspect the approximation level 과 analysis of high dimensional data for visualization 를 위해, interactive visual system 이 개발됐다.
    • they investiated the interactive visual modification within high dimensional data, without compromising the visual quality and analysis.
    • 그들의 interactive visual system 이 실세계 문제에서 유익하고, high dimensional info 의 analysis 에 대해 very supportive 함을 강조했다.
  • [99] 는 visual minterativity 와 data volumns 를 link 하기 위한 challenge 에 worked on 했다.
    • approximate query processing 을 위해 sample-based model 을 조사했다.
    • interactive visualization 을 위한 approximate solution 을 build 하기 위해 explored 됐던 거였다.
    • 해당 모델에서, 알고리즘은 the correctness and approximate answers of query approximation automatically 를 obsert 하기 위해 employ on linaer and non-linear functions 를 사용했다.
  • [100]

@uhhyunjoo
Copy link
Owner Author

B. RECOMMENDATION VISUAL ANALYSIS

image

  • 이러한 추천 시스템의 목적 : 데이터의 숨겨진 의미와 값을 자동적으로 들춰내는 것이다.

  • effective and interactive 한 (이하 좋은) 차트, 그래프, 맵을 추천하는 다양한 알고맂므들은 제안되어 왔다.

  • statistical 한 techniques 를 사용해서, error 와 그에 대한 solution 을 강조하는 추천 알고리즘들 또한 제안되어 왔다.

  • [102] : 경험적/이론적 데이터시각화를 위한 visual 알고리즘을 비교하는 프레임워크 제안

    • 좋은 시각화를 위한 추천을 만드는, central connecting system 에 대해 작동하는 프레임워크다.
    • 좋은 시각화를 위한 best space 를 리스트업하고 랭크를 메긴다. comparing and approximating 을 함으로써.
    • 세 가지 구성요소가 있다
        1. 효과적인 시각화를 위한 네트워크
        1. 추천 시스템을 위한 알고리즘
        1. 리스트업될 시각화에 대한 랭킹과 approximation 을 위한 prediction,
  • [103] : 데이터시각화를 위한 rule-based 추천 시스템

    • 해당 논문에서, 저자들은 knowledge-based rule 을 만들었다. 좋은 데이터 시각화를 위해.
    • 제안된 추천 시스템은 유저 피드백, 의도된 태스크 와 같은 특성을 사용한다.
    • 아래와 같은 컨트리뷰션
        1. Visual analysis 를 위한 data classification taxonomy
        1. mapping visual structures 를 위한 Taxonomy
        1. knowledge-based rule 을 위한 Draft rules
  • [104] : Tabular data 에 대한 visual analysis 를 위한 recommender system 을 제안함.

    • 해당 시스템은 bar charts, line, pie, networks 등에 대한 좋은 시각화를 predict 할 수 있다.
    • 제안된 시스템은 intent features 와 data columns 를 시각화 하기 위해, statistical 알고리즘을 사용한다.
    • 저자들은 또한 header를 사용하여 대상 시각화의 중요한 부분을 따라 필수 그래프 col을 찾고 분류하는 데 사용되는 모델을 제안했다.
    • col 과 타겟 데이터에 대해 더 좋은 시각화를 성취하기 위해 neural 알고리즘을 모델에 사용헀다.
  • [105] : personalized visual recommendation 에 대한 issue 에 집중했고, best solution 을 제공하기 윟나 a learning frame work 를 제안했다.

  • 좋은 시각화를 위한 개인 유저의 visual interaciton 에 집중

  • 사용자의 경험에 따라 연관된 시각화를 학습할 수 있다. 시각화가 데이터를 토대로 만들어짐에도 불구하고.

  • [106] : automatic visual an infographic recommentaion 에 대한 survey articel

    • This article reviewed automatic recommendation systems and classfied the visual system into the following visualizations categories such as annotation, graph, network, graph-network and storytelling.
    • The current challenges and their future direction are also reviewed.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant