AN APPROACH TO THE SUPPORTING SYSTEM FOR EVENT PLANNING- TOWARD A KNOWLEDGE RECONSTRUCTION ENGINE -

Shigeki AMITANI, Mikihiko MORI and Koichi HORI

School of Engineering, the University of Tokyo4th bldg., room No.513, Research Centerfor Advanced Science and Technology (RCAST),4-6-1, Komaba, Meguro-ku, Tokyo, 153-8904, JAPAN

ABSTRACT

The main goal of our research is to establish “a methodol-ogyforinvestigationofan event” and toconstruct“a knowl-edge reconstruction system” for designing effective and at-tractive events. The methodology described here is to ar-ticulate gaps between event planners’ intention and eventvisitors’ mental impression from real event sites. The visi-tors’interactions withevent objects were observed and theirbehavior and verbal reports (protocol data) were recordedwithwearable computers and a video camera, incorporationwith Dentsu Inc. Their cognitive processes were analyzedwith the “retrospective report method” of protocol analy-sis. It is innovative to apply wearable computers and pro-tocol analysis to real world problems to investigate human-environment interaction to grasp their cognitive processesmicroscopically in a real event site. In this paper, we aregoingtodescribe themethodologyadoptedforinvestigatingtheir mental transition processes (Visitors’ Mental Transi-tion: VMT) in the real world and some examples of ob-tained results. From analysis of the data, we obtained aprospect that our microscopic and detailed approach is use-ful and effective toward actual event plannins. The conceptand prototype of “knowledge reconstruction system” is de-scribed.

1. INTRODUCTION

Every year, event plannin companies hold various events.So far event plannin is conducted with implicit knowledgeof experienced planners and effectiveness of events is mea-sured only by questionnaires. In actual situations, it is saidthat planners cannot obtain adequate and proper knowledgefor future planning from statistical data derived from ques-tionnaires because:

The questions are composed of only what the plan-ners pay attentionto in advance. That means they failto obtain something unexpected.

Thanks to Dentsu Inc. for cooperation and funding.

The plannerscannotunderstandhowandwhythedatawas produced. It makes it dif

ﬁ

cult for the planners toutilize the knowledge.

Even though free-answer questionnaires are adopted,at real event sites it is dif

ﬁ

cult to let the visitors an-swer the questions in detail.That means planners are unable to evaluate events theydesign. They need to knowhow visitors toan event actuallyfeel and how they behave when they are at the event boothin order to construct strategies for next event plannin. Thisis because knowledge cannot be utilized in real situationsunless it has a context where the knowledge was produced[1]. Though quantitative data is powerful to form a con-clusion, it should be accompanied by qualitative data as areason so that planners can utilize the knowledge. Qualita-tivedataand quantitativedata shouldcomplement witheachother.Inthispaper, weare goingtopropose “a methodologytograsp visitors’ mental transitions” and “a knowledge recon-struction system” through applying the framework of cre-ativity support to actual event designing works. As its

ﬁ

rststep, an experiment was conducted at two actual events, -“Tokyo Motor Show 2001” and “World PC Expo 2001”,in cooperation with Dentsu Inc. The visitors’ interactionswiththe event objects were observed and their behavior andverbal reports (protocol data) were recorded. The conceptandprototypeof“knowledgereconstructionsystem” are de-scribed.

2. METHODOLOGY AND EXPERIMENT

As the aim of this investigation is to grasp visitors’ men-tal transition (VMT) in a microscopic way, we adopted themethodology of “protocol analysis [2]”. To obtain their ac-tual cognitive processes at event sites, the experiment wasconducted at two events: World PC Expo 2001 (WPC: 19-22,Sept., 2001)andTokyoMotorShow2001(MotorShow:

26, Oct. - 7, Nov., 2001) held at Makuhari Messe in Japan.Three booths for WPC and one booth for the Motor Showwere selected for the experiment in cooperation with eventorganizers of the booths.

2.1. Articulationof Planners’ Intention

To investigate what planners intend to convey to visitorsand how they designed a booth to express their intentions,we had interviews with the planners in advance. Planningpapers were obtained in advance to extract messages andthemes that the planners try to convey to visitors. The ques-tionsasked totheplannersare basicallysame: “howdidyoudesign the booth to express your concepts?”

2.2. Recording Units for Collecting Protocol Data

To collect protocol data, two wearable computers were pre-pared

. Because of sponsors’ intention, a normal digitalvideo camera was adopted for collecting most of the pro-tocol data. But this experiment de

ﬁ

nitely shows validityof real-world application of wearable computers. In factDentsu Inc. adopted a wearable computer for other investi-gation in marketing

ﬁ

eld.9 subjects (one person for 3 sessions + one pair for 3sessions) at WPC and 12 subjects (one person for each ses-sion) at Motor Show were employed. Subjects are asked tolook around designated booth(s). After visiting the booth, adetailed interview was conducted. Its procedure is:1.

Retrospective reports with visual aid

: The subjectswere asked to report “what you look at”, “what youthink about it” and “what you do” along with VTRthey recorded as a memory aid. This interview iscalled “VA (Visual Aid) Interview”.2.

Questionsaboutthesubjects’ impression onthe event objects

: The questions were made based on the in-terviews with the planners and the planning papers.This is to investigate how the planners’ intention andthe visitors’ impression match or mismatch witheachother.3.

Keyword questionnaires

: A keyword list was madealso based on theinterviewswiththe planners andtheplanning papers. This is toinvestigate whatkeywordstheplannerspresentedwereimpressive andwhichkey-words the subjects could remember.

In cooperation with MIT Media Lab and Intelligent Cooperative Sys-tems Laboratory at Research Center for Advanced Science and Technol-ogy, the University of Tokyo.

3. RESULTS AND DISCUSSION

Though there are a lot of

ﬁ

ndings that surprised the plan-ners, in this paper we are going to show two examples of a lot of unexpected

ﬁ

ndings that are beyond the planners’expectation at Motor Show. These

ﬁ

ndings devote to createnew knowledge, which is “knowledge reconstruction”.

3.1. Effect of the Other Visitors

The other visitors can provide “a context” that raises degreeof satisfaction of visitor. Following report was obtained:

A companion took a picture with a family. Both of thecompanion and the child smiled. My (= the subject’s) chil-dren also like cars. They would be delighted if I took themhere. That is a good idea.

This observed data was reported to one of the plannersand he hit upon a new strategy:

By inviting families that are customers of the company,the other visitors willfeel in a way mentioned above. More-over, the invitedfamilywillalsofeel better because they feel“they are invited as special guests” and this family can en- joy being a customer of the company, which will be great bene

ﬁ

t to the company, too.

This is a good example of “knowledge reconstruction”.We call it “reconstruction” because implicitly they mightknow:

A visitoris affected by anothervisitor at an event site.

Customers are delighted if they are invited as specialguests.

If customers like the company, it is bene

ﬁ

cial to thecompany.Providing a real context connects these pieces of infor-mation. That means knowledge with “its real context” ob-tained through this analysis can support event-planning if itis properly reported to planners.

3.2. GapabouttheStagebetween Planners andVisitors

Normally event planners agree that length of a main stageshowshouldbe withinten minutes. Althoughlengthofvisi-tors’stayingata boothis usuallyinvestigatedwithquantita-tiveanalyses, thelengthofstayingatastage isleftunchecked.It is unchecked at which scene of a stage visitors leave, ei-ther. That means strategies of designing stages does not ex-istexplicitly

. Soitis worthwhileinvestigatingthelengthof visitors’ stayingand at whichscene they leave there. An ex-ample of analysis of a stage is described. One of the stageswas composed of six scenes with technical information and

Perhaps planners might havestrategies implicitly.

images of bene

ﬁ

t. This stage lasted for twelve minutes andstarted every ten minutes.The planners’ hypotheses about the stage are:

Twelve-minute stage is short enough to let visitorswatch whole of it.

The technical parts of the stage are not so attractivethat it is dif

ﬁ

cult to keep visitors watch these parts.Butfromtheanalysis, suchdata andtendency as followswere obtained:

Average length of stay is about 3.6 min.

The visitors leave or get tired of the stage when theimage of bene

ﬁ

ts is started.Though there was one subject who watched whole of the stage, he reported that he could not understand what thescenes with comfortable images tried to convey. There alsoexists a gap between the planners’ knowledge and actualsituation. This phenomenon also must be reported.These facts might not sound surprising very much, butthese pieces of information have not been reported or de-veloped to “useful knowledge” at real event plannin sites sofar because implicitinformation lacks an importantelementfor utilization, i.e., “a real context” [1]. Our methodologyof microscopic analysis can de

ﬁ

nitely provide the plannerswith“knowledge withitsreal context”. Based onthese factsand analysis, we are going to propose a supporting systemthat provides “knowledge with a real context” with plan-ners. In the next section, the system image is described.

4. TOWARD A KNOWLEDGERECONSTRUCTIONSYSTEM4.1. Concept of the System

Though importance of knowledge has been claimed sincemiddle of the 1980s, main concern of business theory hasbeen how to obtain and accumulate established knowledge.Littleresearch hasbeenconductedonhowinnovative knowl-edge can be created. Nonaka [3] has claimed that there arefour modes of knowledge transition: Socialization, Exter-nalization,CombinationandInternalization. Healsoclaimedthat knowledge should be regarded as spiral-up cyclic pro-cess. Though a lot of companies have attempted to applythis theory to their actual works, it has not been successful.Nonaka’s theory was proposed as a theory for “transitionof knowledge mode”, not for “manipulation of knowledgemode”.Asourinvestigationexternalized alotofknowledgesuc-cessfully, we are going to propose a supporting system toutilize the knowledge. Our approach supports each mode of Nonaka’s theory in a following way. In a later section weare going to describe what the system is like in more detail.

Externalization

This mode is supported by the methodol-ogy we proposed in the former sections. In addition,it is necessary for a user to have “ a browser” to dis-cover unexpected knowledge that was unobtainableor discarded by traditional methods. We are going topropose an interactive browser that promotes a user’sre

ﬂ

ective thinking.

Combination

From the data obtainedthroughthe analysis,this mode is supported by “exhaustive search” and“interaction with spatial representation”. Based onthe conclusion of Hori [4] and Yamamoto et al. [5],spatial representation is adopted to show what werefocused by both the visitors and the planners and re-lationships among the focused objects.

Internalization

It is expected that this system promote theuser’s understanding

Socialization

For example, “pursuasion of clients” is ex-pected by using this system.We adoptspatialrepresentationtopresentdata toa plan-ner. Hori [4] conducted the experiment on effects of spatialrepresentation in a conceptual design. His system namedAA1presentswordsthatrepresentsconcepts theuservaguelyconceives, on a two-dimensional space, the user can changetheirlocationsof the words on the space. This action allowsthe user to clarify gradually his/her concept and such phe-nomenon was observed as the user came to generate newconcept by looking at a blank area on the space. This resultcan be used for our aim.Our system aims to “be used for a real world problem”.Knowledge cannot be separated from actual contexts to uti-lizetheknowledge,i.e., knowledge management. Thatmeansthat it is necessary to preserve knowledge together with “itsreal context”. Event planners need to know how knowl-edge was produced. It is necessary for planners to obtainand understand knowledge, to integrate and create innova-tive knowledge, and to apply knowledge to actual event-planning. And from the event, new knowledge can be ex-tracted again. This “loop” is what we call “knowledge re-construction”. Knowledge must be dynamically integrated,innovated and applied to real world. We are going to con-struct a system for supporting to promote this cyclic pro-cess. Infollowing sections, expected interactionand systemimage are described.

4.2. Expected Interaction withthe System

Snap shots of the system are shown in Figure 1 and 2. Thesystem is composed of two components: “ChronoSpace”and “ContextMap”.Following scenario is an expected interaction with thesystem. At the current state, some of the expected functions