Exploring Visualization of Software-Related data in Extended Reality - Thesis dissertation

Slide 1

Slide 1 text

Exploring Visualization of Software-Related data in Extended Reality David Moreno Lumbreras Directores: Jesús M. González Barahona Gregorio Robles Martínez

Slide 2

Slide 2 text

Context Research Goals Introduction

Slide 3

Slide 3 text

Context Data visualization - Software visualization The city metaphor in data and software visualization. Virtual Reality (VR) Web environment /context

Slide 4

Slide 4 text

Context Research Goals Introduction

Slide 5

Slide 5 text

Research goals Study of the state of the art, focusing on the use of the city metaphor. Create a platform for web-based data visualizations in the XR environment. Development of a web-based version of the city metaphor visualization. Conduct empirical studies for validating BabiaXR-visualizations. Enhance the city metaphor visualization. /goals

Slide 6

Slide 6 text

Systematic Mapping Study State of the art

Slide 7

Slide 7 text

Our research on the literature From 3D to Virtual Reality (VR). The city, and other metaphors in softvis. Visualizing dependencies. Literature reviews in software visualization. /state_of_art

Slide 8

Slide 8 text

Understanding the history of the city metaphor in SoftVis through a SMS /city_metaphor

Slide 9

Slide 9 text

Research Questions RQ1: “Has the use of the city metaphor in software visualization increased?” RQ2: “How does the number of research teams evolve over time?” RQ3: “Are researchers developing more tools that support the city metaphor?” RQ4: “Are the prototypes exploring new emerging immersive technologies?” /RQs Publications Tools

Slide 10

Slide 10 text

Search Strategy and Quality Papers which cite two of the main "3D city software visualization" papers. - C. Knight and M. Munro, "Virtual but visible software," -> First paper talking about visualizing software as cities - R. Wettel and M. Lanza, "Visualizing Software Systems as Cities," -> CodeCity first/main paper Four studies from a 2011 Journal paper, 2 VISSOFT 2020 papers and 1 VISSOFT 2022 paper. Cites of those papers included in the dataset. /strategy

Slide 11

Slide 11 text

Filtering and Categorizing No 2D viz papers, no English, only papers… 169 studies Categorizing studies: 1. Categories selection done by three researchers (METAPHORTYPE, NAME, SOURCE, etc…) 2. Two researchers categorized studies 3. Third researcher analyzed the cases in which there was conflict. /ﬁlters

Slide 12

Slide 12 text

Summary of Results /results

Slide 13

Slide 13 text

/results

Slide 14

Slide 14 text

/results

Slide 15

Slide 15 text

/results

Slide 16

Slide 16 text

/results

Slide 17

Slide 17 text

Papers The use of the city metaphor and its derivatives in software visualization continues to be an active metaphor in the community. Greater technology accessibility fuels software visualization research growth.

Slide 18

Slide 18 text

Tools Advancements in technology = software development more accessible = more approaches in SoftVis. Growing trend in the use of XR in SoftVis but needs still more research.

Slide 19

Slide 19 text

BabiaXR Facilitating Experiments Software Visualization

Slide 20

Slide 20 text

Requirements Resource intensive world Dataviz Web environment Open Source /softvis

Slide 21

Slide 21 text

No content

Slide 22

Slide 22 text

https://tinyurl.com/babiabars

Slide 23

Slide 23 text

https://tinyurl.com/babiamuseum

Slide 24

Slide 24 text

/a-frame https://aframe.io

Slide 25

Slide 25 text

babia-queryjson / babia-querycsv babia-queryes babia-querygithub /queriers

Slide 26

Slide 26 text

babia-ﬁlter babia-treebuilder /data_management

Slide 27

Slide 27 text

1. Pie 2. Doughnut 3. 3D/2D bars 4. 3D/2D cylinders 5. Bubbles 6. Network . . /charts

Slide 28

Slide 28 text

1. Pie 2. Doughnut 3. 3D/2D bars 4. 3D/2D cylinders 5. Bubbles 6. Network . . City /charts

Slide 29

Slide 29 text

Examples babiaxr-vismapper (maps data properties to visualization properties) https://babiaxr.gitlab.io/aframe-babia-components/

Slide 30

Slide 30 text

City https://tinyurl.com/babiacitybasic /city

Slide 31

Slide 31 text

/spiral_treemap

Slide 32

Slide 32 text

BabiaXR Facilitating Experiments Software Visualization

Slide 33

Slide 33 text

Needed to facilitate experiments processes After our feedback and state of the art analysis. Automate processes. Preparation - Conduction - Analysis /feedback

Slide 34

Slide 34 text

New set of components Automatic Environment Creation Auto Camera Conﬁguration Browser Accessibility Time Measurement Answer Recording Position and Rotation tracking Automatic Task Positioning /experiments

Slide 35

Slide 35 text

No content

Slide 36

Slide 36 text

CodeCity Dashboards Virtual Reality vs On-Screen

Slide 37

Slide 37 text

CodeCity: On-screen or in VR? /comparison CodeCity as a 3D visualization - VR better to visualize 3D Comparison of CodeCity in two environments. Software comprehension tasks. Can VR improve software comprehension using the city metaphor for visualizing source code metrics? Know advantages/limitations of VR in this context.

Slide 38

Slide 38 text

Research Questions RQ1: “How does the accuracy of participants immersed in VR compare to that of participants using the on-screen version of BabiaXR-CodeCity?” RQ2: “How does the efﬁciency of participants immersed in VR compare to that of participants using the on-screen version of BabiaXR-CodeCity?” /RQs

Slide 39

Slide 39 text

Environment /environment Simulated 3D on a PC vs. immersive VR on Meta Quest 2 tinyurl.com/babiacodecity

Slide 40

Slide 40 text

Experiment /experiment 24 participants (Industry and Academia) Explanation + Demographic Survey Tasks Feedback Questions

Slide 41

Slide 41 text

Correctness /results

Slide 42

Slide 42 text

Time completion /results

Slide 43

Slide 43 text

Summary + Next steps VR shorter completion time for the tasks, but less accuracy. In Top 5, always over 90%. Validate with larger samples Feedback = Improvement /discussion

Slide 44

Slide 44 text

Second experiment - extension /extension Good feedback from first experiment. Invited for an extension. Repeat the experiment with some changes from the feedback. Compare results and confirm the ones extracted in the first experiment.

Slide 45

Slide 45 text

Changes /changes Better time measurement - talks aloud when starting and tasks displayed on the scene. Location of the tasks in order to avoid bias. Training with the city and non-usable data. Remove the environment on a old release. Add two new tasks to improve the results. 26 new participants

Slide 46

Slide 46 text

Summary VR CodeCity is comparable to on-screen CodeCity. = Correctness ++Efﬁciency. The second experiment, better in terms of reliability of results conﬁrms that.

Slide 47

Slide 47 text

CodeCity Dashboards Virtual Reality vs On-Screen

Slide 48

Slide 48 text

Comparing web dashboards vs BabiaXR dashboards /goal Bitergia - Knowledge of software analytics tools - Data + dashboard. First assessment of Babia dashboards. Comparing them to an on-screen dashboards platform. Are BabiaXR dashboards at least as good as Kibana dashboards for conducting software development analytics tasks?

Slide 49

Slide 49 text

Idea of register the methodology 1. Ensure the methodology is correct. Difﬁcult of publish these studies. 2. Present a Registered Report - ESEM 2021 3. Address comments in order to improve the study 4. Conduct the empirical study 5. Publish the results - On going /pre-experiment

Slide 50

Slide 50 text

Replicate dashboards Same Kibana dashboards on BabiaXR Software development processes data - PRs and Issues Two on-screen dashboards in one VR dashboard Same data /experiment

Slide 51

Slide 51 text

No content

Slide 52

Slide 52 text

tinyurl.com/babiadashboard

Slide 53

Slide 53 text

Research Questions RQ: “Is comprehension of software development processes, via the visualization of their metrics, at least as good as in 2D screens when presented in VR scenes?”. RQ1: “Do the answers obtained in VR provide similar correctness compared to those obtained on-screen?” RQ2: “Do the answers obtained in VR provide similar time to completion compared to those obtained on-screen?” /RQs

Slide 54

Slide 54 text

Experiment 1. Divide 32 participants in 4 groups 1.1. Starting in Kibana and project A, then BabiaXR and project B 1.2. Starting in Kibana and project B, then BabiaXR and project A 1.3. Starting in BabiaXR and project A, then Kibana and project B 1.4. Starting in BabiaXR and project B, then Kibana and project C 2. Training 2.1. Both Kibana and VR 2.2. Data from a non-usable project C 3. Demographics survey 4. Tasks, repeating them in both environment 5. Feedback Questions /RQs

Slide 55

Slide 55 text

We run an statistical analysis for the correctness and completion time Mixed linear method Cohen’s effect size … Find if the difference between groups and On-Screen-VR is statistically important. Main Results /results

Slide 56

Slide 56 text

Summary Found that BabiaXR dashboards are at least as useful in terms of correctness and completion time as Kibana dashboards.

Slide 57

Slide 57 text

Discussion Habit factor on on-screen Repeating tasks = second round better results Isolated VR environment - more focused Tasks favoring VR or a project? Interactions … /discussion

Slide 58

Slide 58 text

Time evolution Dependencies City Extending the city metaphor viz

Slide 59

Slide 59 text

/layout_algorithm Challenges in representing software evolution using treemaps Difﬁculty of adding or removing items in treemaps Layout recalculations and impact on visualization when changes occur Limitations of treemaps for representing software evolution Treemaps as layout

Slide 60

Slide 60 text

/spiral Beneﬁts of the spiral algorithm in accommodating changes in the number of elements Scalability and ﬂexibility of the spiral algorithm for representing software evolution Spiral algorithm for the layout

Slide 61

Slide 61 text

Example

Slide 62

Slide 62 text

No content

Slide 63

Slide 63 text

No content

Slide 64

Slide 64 text

/navigation

Slide 65

Slide 65 text

/live_demo tinyurl.com/babiatime

Slide 66

Slide 66 text

Validation (future) Through comparison with existing tools and metrics visualization methods. User studies with software development practitioners for feedback. Reﬁnement and optimization of the methodology. Investigation of the effectiveness of the spiral algorithm compared to other layout algorithms.

Slide 67

Slide 67 text

Time evolution Dependencies City Extending the city metaphor viz

Slide 68

Slide 68 text

Exploring dependencies data Visualizing software dependencies can provide a powerful way to understand and manage the complexity of modern software systems. NPM ecosystem - NodeJS web applications Graph as common visualization City visualization, elevating quarters for representing dependencies levels /exploring

Slide 69

Slide 69 text

tinyurl.com/babiavissoft23

Slide 70

Slide 70 text

Dependencies level /levels

Slide 71

Slide 71 text

Building location /quarters

Slide 72

Slide 72 text

Sub-dependencies /subdependencies

Slide 73

Slide 73 text

Changing metrics on the ﬂy /metrics

Slide 74

Slide 74 text

Repeated dependencies /repeated

Slide 75

Slide 75 text

Repeated dependencies /replicas

Slide 76

Slide 76 text

license: type of license of the package (categoric). timesInstalled: how many times the package is installed, that means, in how many differents versions the package is in the project. timesAppear: how many times the package appears in the project regardless the version. last_act_days: days since the last commit in the repository of the package. ncommits_ly: number of commits in the repository of the package in the last year. ncommiters_ly: number of unique commiters in the repository of the package in the last year. nvuln: number of vulnerabilities of the package, retrieved using the audit tool of npm. nissues_ratio: number of issues closed divided by total number of issues. Building metrics /color

Slide 77

Slide 77 text

Views License license as color Replication timesInstalled or timesAppear as color Activity last_act_days or ncommits_ly or ncommiters_ly as color Vulnerabilities nvuln or nissues_ratio as color /views

Slide 78

Slide 78 text

First feedback /feedback 4 participants from industry Training Open conversation Feedback questions

Slide 79

Slide 79 text

Results License: project policies, decision. Replication: core deps, performance. Activity: maintenance priority. Vulnerabilities: security and stability. By providing multiple views and metrics, developers can gain a more comprehensive understanding of their software dependencies and identify potential issues before they become major problems.

Slide 80

Slide 80 text

Contributions Future Research Conclusion

Slide 81

Slide 81 text

Research contributions Systematic Mapping Study. Development of BabiaXR. Facilitating VR data visualization experiments. Experiment on BabiaXR-CodeCity. Comparison of Kibana and BabiaXR using the museum metaphor. Enhanced CodeCity with Time Evolution Dependencies ecosystem visualization. /contributions

Slide 82

Slide 82 text

Our journey: SMS + Development + Validation + Enhancements = Path for future advancements. Softvis in immersive web environments. Our contributions wanted to have impact on the software visualization research communities. Conclusions

Slide 83

Slide 83 text

Contributions Future Research Conclusion

Slide 84

Slide 84 text

Future Research Validation of the Time Evolution Feature. Empirical Study of the Elevated City. Exploration of Augmented Reality (AR) in Software Visualization. Analysis of Alternative Metaphors in VR. General Platform Improvement and Ongoing Research. /future

Slide 85

Slide 85 text

Replicability of the results /replicability https://thesis-dlumbrer.gitlab.io

Slide 86

Slide 86 text

BabiaXR contributing Main page https://babiaxr.gitlab.io Tutorials, APIs, Source Code, etc. GPLv3 License Contributing guide https://gitlab.com/babiaxr/aframe-babia-components/-/blob/master/docs/CONTRIBUTING.md /open_source

Slide 87

Slide 87 text

My PhD timeline PhD starts Jan’19 2019 2020 2021 2022 2023 Research stay @ Software Institute - USI Aug’22 - Nov’22 IST paper accepted Sep’22 Defense Oct’23 VISSOFT Oct’23 Technical Paper Artifact (Best artifact award) VISSOFT Sep’21 Technical Paper Artifact Remotely collaboration starts w/ Michele Lanza Sep’20 Industrial PhD @ Bitergia Jan’19 - Mar’22 Comunidad de Madrid grant Academia @ URJC Mar’22 Pre-doctoral Researcher Journal submission Jan’23 Status: 3rd revision Journal submission Jul’22 Status: 2nd revision Journal submission Apr’23 Status: Major revision ESEM Apr’23 Registered Report MADSESE Jun’19 First publication BENEVOL Nov’19 Workshop paper BENEVOL Sep’22 Workshop paper OpenVRLab Mar’22 Workshop paper SATToSE Jun’23 Workshop paper

Slide 88

Slide 88 text

Exploring Visualization of Software-Related data in Extended Reality David Moreno Lumbreras Directores: Jesús M. González Barahona Gregorio Robles Martínez