Se ha denunciado esta presentación.
Utilizamos tu perfil de LinkedIn y tus datos de actividad para personalizar los anuncios y mostrarte publicidad más relevante. Puedes cambiar tus preferencias de publicidad en cualquier momento.

Supporting job mediator and job seeker through an actionable dashboard

Job mediation services can assist job seekers in finding suitable employment through a personalised approach. Consultation or mediation sessions, supported by personal profile data of the job seeker, help job mediators understand personal situation and requests. Prediction and recommendation systems can directly provide job seekers with possible job vacancies. However, incorrect or unrealistic suggestions, and bad interpretations can result in bad decisions or demotivation of the job seeker. This paper explores how an interactive dashboard visualising prediction and recommendation output can help support the dialogue between job mediator and job seeker, by increasing the "explainability" and providing mediators with control over the information that is shown to job seekers.

  • Inicia sesión para ver los comentarios

  • Sé el primero en recomendar esto

Supporting job mediator and job seeker through an actionable dashboard

  1. 1. Supporting job mediator and job seeker through an actionable dashboard Francisco Gutiérrez, Sven Charleer & Katrien Verbert https://augment.cs.kuleuven.be/ 1
  2. 2. [background] !2
  3. 3. [background] !3
  4. 4. [background] !4 Job Seeker Mediation Sessions “Dream Job”
  5. 5. Services: Register in the service. Update profile. Advising job seekers who struggle to find a job. 5 [context] VDAB employs job mediators all over the country to support job seekers. Mediation sessions can last from 15 minutes to two hours.
  6. 6. [context] !6 RECOMMENDER SYSTEM Black Box (?) Dream Job MEDIATORJOB SEEKER
  7. 7. 7 [context] The Mediation service wants to improve advice sessions through a new recommender system! Three years of data, 700 000 job seekers. Predicts the chance of finding a job within 120 days.
  8. 8. 8 [context] Original dashboard for mediators Age Studies Experience and competences Languages Education Increase opportunities Teacher Chance of finding the job within 120 days
  9. 9. 9 [context] Original dashboard for mediators
  10. 10. 10 [goal] Design a tool that attempts to increase: Trust Control and justification of predictions Acceptance of recommendations Supportive and collaborative -> (during mediation sessions)
  11. 11. Support the dialogue between “expert” and “laymen” “… the expert can become the intermediary between the [system] and the [end-user] in order to avoid misinterpreta4on and incorrect decisions on behalf of the data… “ Sven Charleer, Andrew Vande Moere, Joris Klerkx, Katrien Verbert, and Tinne De Laet. 2017. Learning Analytics Dashboards to Support Adviser-Student Dialogue. IEEE Transactions on Learning Technologies (2017), 1–12. 11 [goal] Learning analytics, student counseling
  12. 12. 12 [preliminary study] Costumer journey approach During one day workshop to gain insight into typical job seeker-mediator session (n = 5) Observation of hands-on time With the original dashboard Observations of individual mediation sessions Between mediators and job-seekers (Nm = 3, Njs= 6) 15-30 min 5-Likert Scale Questionnaire Perceptions of dashboard and predictions
  13. 13. 13 [findings] Contextual Information Lack of data might lead to unrealistic recommendations. RS can take a [supportive role] during the sessions. Transparency Mediators aware of the “blackbox”. Provide further information through visualisation. Controllability Mediator needs to remain in control. Recommender-output <- [mediator] - > Job Seeker (motivation) Justification Provide visual support to convey the message. [reality check] Job-seeker visual literacy plays an important factor.
  14. 14. 14 [design goals] [DG1] Control the message The mediator filters the information flow to convey a message, and avoid potential demotivation. [DG2] Clarify the recommendations The mediator requires further details about predictions [DG3] Support the mediator The dashboard assists the mediator during the session
  15. 15. 15 [working prototype] The dashboard suggests four jobs. …architect (?)
  16. 16. 16 Age was removed to avoid demotivation. Days unemployed was left as “eye opener”. [working prototype]
  17. 17. 17 Different ways to present parameter data. (age) [design & development] Forest plot Circles chart Barchart
  18. 18. Qualitative evaluation with expert users: (N = 12, 10f, age: M= 40.7, SD = 9.4) 18 [evaluation] Years of experience: (M = 9, SD = 4.3) Six mediators dealt only with higher education job seekers. Four with secondary to higher education. Two with job seekers without technical/professional education. Semi-structured interviews 1) Feedback on parameter visuals. 2) Interaction feedback with the working prototype dashboard.
  19. 19. Feedback on parameter visuals Circles chart, barchart. forest plot out-loud interpretation answer a 5-likert scale regarding clarity of visuals 19 [evaluation] Interaction feedback with prototype Screen recorded, data included errors on purpose. Task: “prepare the dashboard for a costumer that is recently unemployed, experienced in cleaning and wishes to change careers if possible.”
  20. 20. 20 [results] [DG1] Control the message * Five mediators used negative parameters to support their message. * Two mediators removed negative parameters to avoid demotivation. Two themes (1) Customization “Incorrect predictions must go” “age can be demotivating” “too much information might be difficult to process” “would like to see an overview of everything” “depends on the job seeker” (2) Importance of the human factor Data is only part of the information on the job seeker and misses external context gathered during the session.
  21. 21. 21 [results] [DG2] Clarify recommendations Two themes (1) Understanding the visualisation Circles was considered the most clear representation to mediators. Too much information might be hard to interpret. “Can be used as an eye-opener” “If all parameters are negative dashboard should not be shown” (2) Convincing power Certain data might be confrontational and demotivating, customization helps the mediator filter the message. higher education lower education
  22. 22. 22 [results] [DG3] Support the mediator The way of using the dashboard is highly dependent on the mediator and the situation. Useful cases: Orientation cases The job seeker does not have a good idea of possible career paths. The mediator can “help orientate” them. Mediator is stuck Can be used as a “starting point”. provide support to guide long conversations. Deal with “problem cases”, “does not know what he/ she wants and is being resistant.
  23. 23. 23 [conclusions] With our tool experts become “gatekeepers” of the data Dashboard carries the message, mediator “moulds” the message. RS with increased “explainability” increase user’s trust and provides them with actionable insights Customization is used beyond data filtering Visualization supports a dialogue creating “reality checks” by revealing negative factors and assist in motivation.
  24. 24. 24 [conclusions] Job mediators face a broad job-seeker audience. Different educational backgrounds, social situations, personalities. Incompleteness and lack of profile information Makes visual job search and suggestions “as-is” difficult. We attempt to clarify recommendations to both mediators and job seekers. Help mediators to control the message they wish to convey depending on the context.
  25. 25. 25 [future work] Deploy the dashboard in realistic settings! Gain further insight into the impact of the tool in mediation sessions. Labor market exploration. Explore tools to enable job seekers to explore the labor market in a personalized way.
  26. 26. 26 thanks! questions… (?) Sven Charleer Phd. Computer Science Freelance UX researcher/designer sven.charleer@gmail.com @SvenCharleer | Twitter SvenCharleer.com Francisco Gutiérrez Phd. Student Augment HCI, KU Leuven francisco.gutierrez@cs.kuleuven.be @FranciscoGhz | Twitter augment.cs.kuleuven.be

×