pith. sign in

arxiv: 2606.24236 · v1 · pith:MRUIDQH4new · submitted 2026-06-23 · 📊 stat.ML · cs.CV· cs.LG

Automated Residual Plot Assessment With the R Package autovi and the Shiny Application autovi.web

Weihao Li , Dianne Cook , Emi Tanaka , Susan VanderPlas , Klaus Ackermann This is my paper

Pith reviewed 2026-06-25 22:29 UTC · model grok-4.3

classification 📊 stat.ML cs.CVcs.LG
keywords residual plotsmodel diagnosticscomputer visionautomationvisual signal strengthlinear modelsR package
0
0 comments X

The pith

A computer vision model predicts visual signal strength to automate residual plot assessment for linear models.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Visual assessment of residual plots is a standard way to diagnose linear models but it is subjective and does not scale. The paper introduces an R package that applies a computer vision model to a sample of residuals and outputs a visual signal strength score plus supporting details. An accompanying Shiny application makes the tool easy to use. The approach replaces manual inspection with an automated prediction that aims to deliver consistent judgments about model fit.

Core claim

The authors present a computer vision model that, given residuals from a fitted linear model, predicts a visual signal strength value and supplies additional information to help decide whether the model fits the data adequately.

What carries the argument

The computer vision model that outputs the visual signal strength (VSS) score from residual plots.

If this is right

  • Assessments of model fit become consistent across different analysts.
  • Evaluation of residual plots can be performed on large numbers of models without proportional increase in human effort.
  • Analysts receive quantitative scores and supporting details that supplement visual inspection.
  • The method complements protocols that embed observed plots among null plots by reducing the required human judgments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same model architecture could be retrained on other diagnostic plots such as Q-Q plots or leverage plots.
  • Embedding the VSS calculation inside routine modeling functions would allow automatic flagging during data analysis.
  • Repeated application to residuals from many models could reveal systematic patterns in how certain model types produce visual signals.

Load-bearing premise

The visual signal strength score from the computer vision model matches the diagnostic patterns a human analyst would see in residual plots.

What would settle it

Compare the model's VSS scores against ratings collected from multiple human analysts on the same set of residual plots; low agreement between the scores and the human ratings would falsify the claim.

Figures

Figures reproduced from arXiv: 2606.24236 by Dianne Cook, Emi Tanaka, Klaus Ackermann, Susan VanderPlas, Weihao Li.

Figure 1
Figure 1. Figure 1: Reading residual plots can be a difficult task, particularly for students new to [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: True plot alongside one null plot, for quick comparison. [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Summary plot comparing the densities of VSS for bootstrapped residual samples [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Diagram illustrating the infrastructure of the [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: A lineup of residual plots from Poisson generalised linear models, with the true [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: To begin the workflow for autovi using the lineup example dataset, the user clicks the ‘Use Lineup Example’ button (left) to load the example dataset, during which the data status and CSV type will be automatically updated. The user must manually select the label for the true residual plot (middle) to compute further results. The user initiates the assessment of the lineup example data by clicking the run … view at source ↗
Figure 7
Figure 7. Figure 7: Results for the lineup. The VSS of the true residual plot is displayed in the first row [PITH_FULL_IMAGE:figures/full_fig_p018_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Summaries assessing the strength of the pattern and which elements of the plot [PITH_FULL_IMAGE:figures/full_fig_p019_8.png] view at source ↗
read the original abstract

Visual assessment of residual plots is a common approach for diagnosing linear models, but it relies on manual evaluation, which does not scale well and can lead to inconsistent decisions across analysts. The lineup protocol, which embeds the observed plot among null plots, can reduce subjectivity but requires even more human effort. In today's data-driven world, such tasks are well suited for automation. We present a new R package that uses a computer vision model to automate the evaluation of residual plots. An accompanying Shiny application is provided for ease of use. Given a sample of residuals, the model predicts a visual signal strength (VSS) and offers supporting information to help analysts assess model fit.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The paper presents the R package autovi and accompanying Shiny application autovi.web, which use a computer vision model to predict a visual signal strength (VSS) score from a sample of residuals in order to automate the visual assessment of residual plots for diagnosing linear models, thereby addressing the scalability and consistency limitations of manual evaluation and the lineup protocol.

Significance. If the VSS output were shown to reliably track diagnostic patterns such as non-normality, heteroscedasticity, or nonlinearity in a manner consistent with human judgment, the package would offer a practical automation tool that could scale residual-plot diagnostics to large modeling workflows and reduce inter-analyst variability.

major comments (1)
  1. [Abstract] Abstract: The central claim that the computer vision model automates residual-plot assessment by outputting a VSS that helps analysts judge model fit cannot be evaluated, because the manuscript supplies no training details, validation metrics, performance data, or comparison to human judgments or the lineup protocol.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their review and constructive feedback. We address the single major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that the computer vision model automates residual-plot assessment by outputting a VSS that helps analysts judge model fit cannot be evaluated, because the manuscript supplies no training details, validation metrics, performance data, or comparison to human judgments or the lineup protocol.

    Authors: We agree that the current manuscript does not supply the requested details on model training, validation metrics, performance data, or direct comparisons to human judgment or the lineup protocol. The paper is written as a software description focused on the autovi package and autovi.web application. To allow proper evaluation of the central claim regarding the VSS score, we will add a new section (or subsection) describing the computer vision model, including training data and procedure, model architecture, validation metrics, and any available quantitative or qualitative comparisons to human assessments. We will also update the abstract to reflect these additions. This constitutes a major revision. revision: yes

Circularity Check

0 steps flagged

No derivation chain or self-referential steps present

full rationale

The manuscript presents an R package and Shiny app that applies a computer-vision model to residual plots and outputs a visual signal strength (VSS) score. The abstract and available text contain no equations, no parameter-fitting procedure, no uniqueness theorems, and no self-citations that bear the central claim. Because no derivation or prediction step is described that could reduce to its own inputs by construction, none of the enumerated circularity patterns can be exhibited. The work is therefore self-contained against the circularity criteria.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only information prevents identification of any free parameters, axioms, or invented entities; no technical details are supplied to populate the ledger.

pith-pipeline@v0.9.1-grok · 5653 in / 1032 out tokens · 33826 ms · 2026-06-25T22:29:02.225837+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

165 extracted references · 4 canonical work pages

  1. [1]

    Tenth International Conference on Machine Vision (ICMV 2017) , editor =

    Nima Hatami and Yann Gavet and Johan Debayle , title =. Tenth International Conference on Machine Vision (ICMV 2017) , editor =. 2018 , doi =

    2017

  2. [2]

    The American Statistician , volume=

    Diagnostic value of residual and partial residual plots , author=. The American Statistician , volume=. 1987 , publisher=

    1987

  3. [3]

    1982 , publisher=

    Residuals and influence in regression , author=. 1982 , publisher=

    1982

  4. [4]

    1980 , publisher=

    Regression analysis and its application: a data-oriented approach , author=. 1980 , publisher=

    1980

  5. [5]

    Technometrics , volume=

    A graphical technique for enhancing scatterplots with moving statistics , author=. Technometrics , volume=. 1975 , publisher=

    1975

  6. [6]

    Technometrics , volume=

    Exploring partial residual plots , author=. Technometrics , volume=. 1993 , publisher=

    1993

  7. [7]

    Statistica Sinica , pages=

    The examination of residual plots , author=. Statistica Sinica , pages=. 1998 , publisher=

    1998

  8. [8]

    1998 , publisher=

    Applied regression analysis , author=. 1998 , publisher=

    1998

  9. [9]

    Technometrics , volume=

    The examination and analysis of residuals , author=. Technometrics , volume=. 1963 , publisher=

    1963

  10. [10]

    1999 , publisher=

    Applied regression including computing and graphics , author=. 1999 , publisher=

    1999

  11. [11]

    1980 , publisher=

    Regression diagnostics: Identifying influential data and sources of collinearity , author=. 1980 , publisher=

    1980

  12. [12]

    1982 , publisher=

    Introduction to linear regression analysis , author=. 1982 , publisher=

    1982

  13. [13]

    Econometrica: Journal of the Econometric Society , pages=

    A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity , author=. Econometrica: Journal of the Econometric Society , pages=. 1980 , publisher=

    1980

  14. [14]

    Econometrica: Journal of the Econometric Society , pages=

    A simple test for heteroscedasticity and random coefficient variation , author=. Econometrica: Journal of the Econometric Society , pages=. 1979 , publisher=

    1979

  15. [15]

    Journal of the Royal Statistical Society: Series B (Methodological) , volume=

    Tests for specification errors in classical linear least-squares regression analysis , author=. Journal of the Royal Statistical Society: Series B (Methodological) , volume=. 1969 , publisher=

    1969

  16. [16]

    Journal of the American Statistical Association , volume=

    Graphical perception: Theory, experimentation, and application to the development of graphical methods , author=. Journal of the American Statistical Association , volume=. 1984 , publisher=

    1984

  17. [17]

    Computational Statistics , volume=

    Using visual statistical inference to better understand random class separations in high dimension, low sample size data , author=. Computational Statistics , volume=. 2015 , publisher=

    2015

  18. [18]

    Joint Statistics Meetings, August , year=

    Inference for data visualization , author=. Joint Statistics Meetings, August , year=

  19. [19]

    Journal of Computational and Graphical Statistics , volume=

    Exploratory data analysis for complex models , author=. Journal of Computational and Graphical Statistics , volume=. 2004 , publisher=

    2004

  20. [20]

    International Statistical Review , volume=

    A Bayesian formulation of exploratory data analysis and goodness-of-fit testing , author=. International Statistical Review , volume=. 2003 , publisher=

    2003

  21. [21]

    and Carlin, J

    Gelman, A. and Carlin, J. B. and Stern, H. S. and Dunson, D. B. and Vehtari, A. and Rubin, D. B. , year=2013, title=

    2013

  22. [22]

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , volume=

    Statistical inference for exploratory data analysis and model diagnostics , author=. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , volume=. 2009 , publisher=

    2009

  23. [23]

    Journal of the American Statistical Association , volume=

    Validation of visual statistical inference, applied to linear models , author=. Journal of the American Statistical Association , volume=. 2013 , publisher=

    2013

  24. [24]

    IEEE Transactions on Visualization and Computer Graphics , volume=

    Graphical tests for power comparison of competing designs , author=. IEEE Transactions on Visualization and Computer Graphics , volume=. 2012 , publisher=

    2012

  25. [25]

    The American Statistician , volume=

    Variations of Q--Q plots: the power of our eyes! , author=. The American Statistician , volume=. 2016 , publisher=

    2016

  26. [26]

    The Annals of Mathematical Statistics , volume=

    The Lagrangian multiplier test , author=. The Annals of Mathematical Statistics , volume=. 1959 , publisher=

    1959

  27. [27]

    Stat , volume=

    Statistical significance calculations for scenarios in visual inference , author=. Stat , volume=. 2021 , publisher=

    2021

  28. [28]

    arXiv preprint arXiv:1408.1974 , year=

    Human factors influencing visual statistical inference , author=. arXiv preprint arXiv:1408.1974 , year=

    Pith/arXiv arXiv 1974

  29. [29]

    Wiley Interdisciplinary Reviews: Computational Statistics , volume=

    Diagnostic tools for hierarchical linear models , author=. Wiley Interdisciplinary Reviews: Computational Statistics , volume=. 2013 , publisher=

    2013

  30. [30]

    Journal of the American Statistical Association , volume=

    Science and statistics , author=. Journal of the American Statistical Association , volume=. 1976 , publisher=

    1976

  31. [31]

    Biometrika , volume=

    An analysis of variance test for normality (complete samples) , author=. Biometrika , volume=. 1965 , publisher=

    1965

  32. [32]

    Economics Letters , volume=

    Efficient tests for normality, homoscedasticity and serial independence of regression residuals , author=. Economics Letters , volume=. 1980 , publisher=

    1980

  33. [33]

    Journal of Computational and Graphical Statistics , volume=

    Are you normal? The problem of confounded residual structures in hierarchical linear models , author=. Journal of Computational and Graphical Statistics , volume=. 2015 , publisher=

    2015

  34. [34]

    Journal of Statistical Software , volume=

    HLMdiag: A suite of diagnostics for hierarchical linear models in R , author=. Journal of Statistical Software , volume=

  35. [35]

    ac—A subject pool for online experiments , author=

    Prolific. ac—A subject pool for online experiments , author=. Journal of Behavioral and Experimental Finance , volume=. 2018 , publisher=

    2018

  36. [36]

    1964 , publisher=

    Handbook of mathematical functions with formulas, graphs, and mathematical tables , author=. 1964 , publisher=

    1964

  37. [37]

    Laplace, Pierre-Simon , volume=. Th. 1820 , publisher=

  38. [38]

    Sur un nouveau d

    Hermite, M , year=. Sur un nouveau d

  39. [39]

    The R Journal , year =

    David Kahle , title =. The R Journal , year =

  40. [40]

    R News , year =

    Diagnostic checking in regression relationships , author =. R News , year =

  41. [41]

    2020 , note =

    skedastic: Heteroskedasticity diagnostics for linear regression models , author =. 2020 , note =

    2020

  42. [42]

    Economics Letters , volume=

    A note on algebraic equivalence of White's test and a variation of the Godfrey/Breusch-Pagan test for heteroscedasticity , author=. Economics Letters , volume=. 1983 , publisher=

    1983

  43. [43]

    The Annals of Mathematical Statistics , volume=

    On information and sufficiency , author=. The Annals of Mathematical Statistics , volume=. 1951 , publisher=

    1951

  44. [44]

    2022 , note =

    cli: Helpers for developing command line interfaces , author =. 2022 , note =

    2022

  45. [45]

    2022 , note =

    curl: A modern and flexible web client for R , author =. 2022 , note =

    2022

  46. [46]

    2023 , note =

    dplyr: A grammar of data manipulation , author =. 2023 , note =

    2023

  47. [47]

    arXiv:1403.2805 [stat.CO] , year =

    The jsonlite Package: A practical and consistent mapping between JSON data and R objects , author =. arXiv:1403.2805 [stat.CO] , year =

    Pith/arXiv arXiv

  48. [48]

    2019 , note =

    progress: Terminal progress bars , author =. 2019 , note =

    2019

  49. [49]

    2022 , note =

    tibble: Simple data frames , author =. 2022 , note =

    2022

  50. [50]

    2021 , note =

    ggmosaic: Mosaic plots in the 'ggplot2' framework , author =. 2021 , note =

    2021

  51. [51]

    2022 , note =

    purrr: Functional programming tools , author =. 2022 , note =

    2022

  52. [52]

    2022 , note =

    tidyr: Tidy messy data , author =. 2022 , note =

    2022

  53. [53]

    2022 , note =

    readr: Read rectangular text data , author =. 2022 , note =

    2022

  54. [54]

    2022 , note =

    stringr: Simple, consistent wrappers for common string operations , author =. 2022 , note =

    2022

  55. [55]

    2020 , note =

    here: A simpler way to find your files , author =. 2020 , note =

    2020

  56. [56]

    2021 , note =

    kableExtra: Construct complex table with kable and pipe syntax , author =. 2021 , note =

    2021

  57. [57]

    2022 , note =

    patchwork: The composer of plots , author =. 2022 , note =

    2022

  58. [58]

    2018 , note =

    Jakub Nowosad , title =. 2018 , note =

    2018

  59. [59]

    PythonAnywhere , author=

  60. [60]

    2018 , publisher=

    Flask web development: Developing web applications with Python , author=. 2018 , publisher=

    2018

  61. [61]

    Behavior Research Methods , volume=

    jsPsych: A JavaScript library for creating behavioral experiments in a web browser , author=. Behavior Research Methods , volume=. 2015 , publisher=

    2015

  62. [62]

    2022 , note =

    rticles: Article formats for R Markdown , author =. 2022 , note =

    2022

  63. [63]

    knitr: A comprehensive tool for reproducible research in

    Yihui Xie , publisher =. knitr: A comprehensive tool for reproducible research in. Implementing reproducible computational research , editor =. 2014 , note =

    2014

  64. [64]

    2020 , note =

    R Markdown cookbook , author =. 2020 , note =

    2020

  65. [65]

    Computational Statistics & Data Analysis , volume=

    A study of partial F tests for multiple linear regression models , author=. Computational Statistics & Data Analysis , volume=. 2007 , publisher=

    2007

  66. [66]

    Journal of Statistics and Data Science Education , volume=

    Bringing visual inference to the classroom , author=. Journal of Statistics and Data Science Education , volume=. 2021 , publisher=

    2021

  67. [67]

    Journal of Computational and Graphical Statistics , volume=

    value plots: Boxplots for large data , author=. Journal of Computational and Graphical Statistics , volume=. 2017 , publisher=

    2017

  68. [68]

    BMC Medical Research Methodology , volume=

    The relationship between statistical power and predictor distribution in multilevel logistic regression: a simulation-based approach , author=. BMC Medical Research Methodology , volume=. 2019 , publisher=

    2019

  69. [69]

    2018 , publisher=

    Practical statistical power analysis using Webpower and R , author=. 2018 , publisher=

    2018

  70. [70]

    Econometrica: Journal of the Econometric Society , pages=

    Partial time regressions as compared with individual trends , author=. Econometrica: Journal of the Econometric Society , pages=. 1933 , publisher=

    1933

  71. [71]

    Educational and psychological measurement , volume=

    Practical significance: A concept whose time has come , author=. Educational and psychological measurement , volume=. 1996 , publisher=

    1996

  72. [72]

    2011 , publisher=

    Thinking, fast and slow , author=. 2011 , publisher=

    2011

  73. [73]

    Journal of the American Statistical Association , volume=

    Errors of the third kind in statistical consulting , author=. Journal of the American Statistical Association , volume=. 1957 , publisher=

    1957

  74. [74]

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , author =

    Statistical inference for exploratory data analysis and model diagnostics , volume =. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , author =. 2009 , pages =. doi:10.1098/rsta.2009.0120 , number =

    work page doi:10.1098/rsta.2009.0120 2009

  75. [75]

    Journal of the American Statistical Association , author =

    Validation of visual statistical inference, applied to linear models , volume =. Journal of the American Statistical Association , author =. 2013 , pages =. doi:10.1080/01621459.2013.808157 , number =

    work page doi:10.1080/01621459.2013.808157 2013

  76. [76]

    Visual Mining Methods for RNA-Seq Data: Data Structure, Dispersion Estimation and Significance Testing , journal=

    Tengfei Yin and Mahbubul Majumder and Niladri. Visual Mining Methods for RNA-Seq Data: Data Structure, Dispersion Estimation and Significance Testing , journal=

  77. [77]

    Spatial Reasoning and Data Displays , year=

    VanderPlas, Susan and Hofmann, Heike , journal=. Spatial Reasoning and Data Displays , year=

  78. [78]

    Graphical inference for infovis , year=

    Wickham, Hadley and Cook, Dianne and Hofmann, Heike and Buja, Andreas , journal=. Graphical inference for infovis , year=

  79. [79]

    2015 , month = jun, journal =

    Using Visual Statistical Inference to Better Understand Random Class Separations in High Dimension, Low Sample Size Data , author =. 2015 , month = jun, journal =. doi:10.1007/s00180-014-0534-x , urldate =

    work page doi:10.1007/s00180-014-0534-x 2015

  80. [80]

    Mahbubul Majumder, Tengfei Yin , year =. Visual. Journal of Data Mining in Genomics & Proteomics , volume =. doi:10.4172/2153-0602.1000139 , urldate =

    work page doi:10.4172/2153-0602.1000139

Showing first 80 references.