Page - 330 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

Table1. Translationexamplewith IBMToneAnalyzer Text Anger Disgust Fear Joy Sadness he good thing about hav- ing good partners who knowwhat theydo 0.0218 0.007 0.080 0.744 0.060 Themaincontributionof thispaper is thecreationofanemotionalmap in real time using the dataset proposed in [16]. Thiswill allowus to knowhowemotions expressed in user postedTweets are distributed anddisplayed across a specific geographical area. Each tweet published with the application, whose architecture is shown in the Figure 2(A)will be processedwith ToneAnalyzer and assigned values from each of the five emotions shown inTable 1. The emotional valueswill form the input into a rule based fuzzy logic system proposed in [18], and depected in Figure 1. The fuzzy systemwill map the input emotions to outputs representing twovalues: energy and pleasantness as indicatorsofwellbeing. Figure 1. Fuzzy system to convert the response of 5 IBM data, in Energy and pleassantness acording to RULER’smodel. Imageadapted from[18]. Our methodology also allows the user to captures a Likert self-evaluation of the same emotions generated by IBM, prior to the publication of their Tweet. This can be used tocompare,validateand later self configure the resultsof the system. In section 4,we can see three visual representations of the emotionalmap for dif- ferent areas of the city andduringdifferent times, alongwith their respective frequency tables. 4. Experiments Inorder to evaluate the capabilityof theproposed system tocapture andgeographically map emotions, 30undergraduate students from InstituteThecnological ofLeo´n partici- pated in experiments over a periodof twoweeksusing themobile app,whose architec- ture is shown inFigure 2(A)[16]. In this experiments the participantswere expected to interactwith theappat least3 timesperday,sharing theirmood,activities,andthoughts. Participants in theexperimentweregivenan informative talk inwhichweexplained the reason for the experiment, the operationof the app, and the data it sends and saves. J.J.ManriquezSantosetal. /ACrowdsourcingBasedComputationalMethodology330