Analysis of Beer Can Design Using QT1 based on High_Grade/Ambience Concept
Inayatus
October 29, 2019
1 Setup
In this learn by building, we will try to make Kansei Engineering analysis of Beer Can Design using QT1 based on High Grade and Ambience. First of all, we must import the library that used to make Kansei Engineering and data wrangling.
2 Background
(Japanese: 感性工学 kansei kougaku, emotional / affective engineering) aims at the development or improvement of products and services by translating the customer’s psychological feelings and needs into the domain of product design (i.e. parameters).
The idea of Kansei engineering was conceived in 1970. Since then, Mitsuo Nagamachi, the founder, has worked to establish the methodology of Kansei engineering, and has assisted in developing no less than 50 new products.
The R&D team grasps the customer’s feeling, namely the Kansei; analyzes the Kansei data using psychological, ergonomic, medical, or engineering methods; and designs the new product based on the analyzed information. Kansei/affective engineering is a technological and engineering process from Kansei data into design specifications.
3 Data Preparation
We want to try design a product of Beer Can using Quantification Theory Type 1 method based on Kansei Engineering. Using 56 sample of beer can, we want to try design element that should be exist in designing a beer can. There are Color, Illustration, and Label Shape. Besides, each of the design element has several category. The list of design elements and its category presented below.
table1 <- data.frame(Can_Color = c("1: White", "2: Silver", "3: Gold", "4: Blue", "5: Black",
"6: Red", "7: Green", "8: Cream", "9: Light Blue", "10: Yellow"),
Can_Illustration = c("1: Animal", "2: Bird", "3: Peson", "4: Barley/Hops",
"5: Crown/Symbol", "6: Other Object", "7: No Illustration",
" ", " ", " "),
Label_Shape = c("1: Oval Shape", "2: Other Trad", "3: None of Above",
" "," "," "," "," "," "," " ))
table1Then we must to import the data set of beer can to make the analysis.
By using the beer dataset above, and use High_Grade/Ambience as the target variable.4 Design Modeling Quantification Theory Type 1
To make QT1 model based on high grade and ambience variable.
## [1] "Color" "Illustration" "Shape"## [1] "Color" "Illustration" "Shape"To check the summary of the model, we can use summary() function.
## $coefficients
## category score
## Color.1 -0.237174
## Color.2 -0.805185
## Color.3 0.434745
## Color.4 0.426079
## Color.5 -0.492530
## Color.6 0.452240
## Color.7 0.388780
## Color.8 0.051907
## Color.9 -1.073005
## Color.10 -0.995782
## Illustration.1 -0.386893
## Illustration.2 -0.303600
## Illustration.3 -1.437420
## Illustration.4 0.059631
## Illustration.5 0.320329
## Illustration.6 0.034269
## Illustration.7 -1.009653
## Shape.1 -0.142978
## Shape.2 -0.226828
## Shape.3 0.072140
## constant term 3.310536
##
## $r
## Color Illustration Shape beer$High_Grade
## Color 1.000000 -0.096773 -0.25378 0.52026
## Illustration -0.096773 1.000000 -0.26468 0.49149
## Shape -0.253776 -0.264678 1.00000 -0.17400
## beer$High_Grade 0.520263 0.491494 -0.17400 1.00000
##
## $partial
## partial correlation coefficient t value P value
## Color 0.66975 6.5038 3.0581e-08
## Illustration 0.65318 6.2204 8.6231e-08
## Shape 0.19147 1.4068 1.6544e-01
##
## $prediction
## observed value predicted value residual
## #1 2.75 3.2067 -4.5674e-01
## #2 3.88 4.1552 -2.7524e-01
## #3 3.63 2.9244 7.0559e-01
## #4 3.63 2.8978 7.3218e-01
## #5 4.25 4.1377 1.1225e-01
## #6 2.75 3.4658 -7.1583e-01
## #7 3.25 3.2051 4.4867e-02
## #8 2.38 3.4658 -1.0858e+00
## #9 3.38 2.1906 1.1894e+00
## #10 4.50 3.9140 5.8603e-01
## #11 3.25 3.4559 -2.0594e-01
## #12 3.63 3.9140 -2.8397e-01
## #13 3.75 3.5441 2.0594e-01
## #14 3.50 2.8419 6.5810e-01
## #15 2.50 2.6268 -1.2678e-01
## #16 3.00 3.1798 -1.7977e-01
## #17 3.88 3.8771 2.9487e-03
## #18 2.50 2.9954 -4.9536e-01
## #19 2.50 2.5033 -3.2517e-03
## #20 3.75 3.8771 -1.2705e-01
## #21 4.00 3.9226 7.7368e-02
## #22 3.88 3.4942 3.8579e-01
## #23 4.00 3.7549 2.4509e-01
## #24 3.63 3.4658 1.6417e-01
## #25 1.88 2.6118 -7.3176e-01
## #26 3.38 3.9140 -5.3397e-01
## #27 2.63 2.7991 -1.6910e-01
## #28 4.25 3.2507 9.9929e-01
## #29 2.38 2.3800 0.0000e+00
## #30 3.25 3.8767 -6.2667e-01
## #31 1.88 2.5033 -6.2325e-01
## #32 1.75 2.8978 -1.1478e+00
## #33 3.63 3.2329 3.9710e-01
## #34 4.38 4.1552 2.2476e-01
## #35 2.75 2.7991 -4.9101e-02
## #36 2.75 2.8078 -5.7767e-02
## #37 4.13 3.4658 6.6417e-01
## #38 2.75 3.2507 -5.0071e-01
## #39 2.00 2.0000 -4.4409e-16
## #40 3.63 3.8767 -2.4667e-01
## #41 2.00 2.5033 -5.0325e-01
## #42 3.50 2.7991 7.0090e-01
## #43 4.38 3.9401 4.3987e-01
## #44 3.63 3.8945 -2.6455e-01
## #45 4.00 4.1552 -1.5524e-01
## #46 3.88 3.9401 -6.0127e-02
## #47 4.75 3.8767 8.7333e-01
## #48 2.00 2.4249 -4.2493e-01
## #49 3.63 3.4658 1.6417e-01
## #50 2.38 2.4220 -4.2004e-02
## #51 3.38 3.4658 -8.5831e-02
## #52 4.13 3.2105 9.1953e-01
## #53 2.63 2.6300 0.0000e+00
## #54 3.00 3.5313 -5.3132e-01
## #55 4.38 4.1552 2.2476e-01
## #56 4.13 4.1377 -7.7495e-03
##
## attr(,"class")
## [1] "qt1"Based on the summary model_qt2 above, color shows the most influential mixture composition that affected high grade of beer.
## $coefficients
## category score
## Color.1 -0.1298216
## Color.2 -0.4729192
## Color.3 0.1428625
## Color.4 0.5270983
## Color.5 -0.4581712
## Color.6 0.3485568
## Color.7 -0.0048712
## Color.8 -0.0079294
## Color.9 -0.6980116
## Color.10 -0.9380098
## Illustration.1 -0.2471606
## Illustration.2 -0.2504281
## Illustration.3 -0.5780329
## Illustration.4 0.1864177
## Illustration.5 0.1328412
## Illustration.6 0.2469450
## Illustration.7 -0.6064160
## Shape.1 -0.0049343
## Shape.2 -0.3064418
## Shape.3 0.0182061
## constant term 3.2969643
##
## $r
## Color Illustration Shape beer$Ambience
## Color 1.00000 -0.13971 -0.175434 0.511724
## Illustration -0.13971 1.00000 -0.169248 0.320575
## Shape -0.17543 -0.16925 1.000000 -0.077241
## beer$Ambience 0.51172 0.32058 -0.077241 1.000000
##
## $partial
## partial correlation coefficient t value P value
## Color 0.60113 5.42433 1.5355e-06
## Illustration 0.47330 3.87450 3.0088e-04
## Shape 0.12235 0.88898 3.7811e-01
##
## $prediction
## observed value predicted value residual
## #1 3.88 3.5720 0.308032
## #2 3.75 3.7966 -0.046568
## #3 3.38 3.1039 0.276056
## #4 3.25 2.9751 0.274908
## #5 3.75 3.5909 0.159126
## #6 3.38 3.3182 0.061810
## #7 3.25 3.3718 -0.121767
## #8 2.25 3.3182 -1.068190
## #9 3.25 2.5951 0.654909
## #10 4.25 3.9520 0.298031
## #11 2.75 3.1154 -0.365434
## #12 3.50 3.9520 -0.451969
## #13 4.13 3.7646 0.365434
## #14 3.38 2.9349 0.445079
## #15 2.63 2.9118 -0.281780
## #16 3.00 3.4323 -0.432294
## #17 3.50 3.6445 -0.144451
## #18 2.75 2.9667 -0.216700
## #19 2.50 2.6098 -0.109839
## #20 3.13 3.6445 -0.514451
## #21 3.63 3.5677 0.062266
## #22 4.00 3.4937 0.506341
## #23 4.25 3.4401 0.809918
## #24 3.50 3.3182 0.181810
## #25 2.88 3.0892 -0.209196
## #26 3.13 3.9520 -0.821969
## #27 2.88 3.2359 -0.355853
## #28 4.00 3.2950 0.704950
## #29 2.88 2.8800 0.000000
## #30 3.13 3.4200 -0.290000
## #31 1.88 2.6098 -0.729839
## #32 2.13 2.9751 -0.845092
## #33 3.63 3.3934 0.236574
## #34 3.50 3.7966 -0.296568
## #35 3.50 3.2359 0.264147
## #36 3.50 2.8516 0.648383
## #37 4.13 3.3182 0.811810
## #38 3.25 3.2950 -0.045050
## #39 2.13 2.1300 0.000000
## #40 2.75 3.4200 -0.670000
## #41 2.25 2.6098 -0.359839
## #42 3.63 3.2359 0.394147
## #43 4.00 3.7734 0.226572
## #44 4.00 3.8501 0.149855
## #45 3.63 3.7966 -0.166568
## #46 3.63 3.7734 -0.143428
## #47 4.38 3.4200 0.960000
## #48 1.75 2.7008 -0.950825
## #49 3.25 3.3182 -0.068190
## #50 3.13 3.0055 0.124471
## #51 3.13 3.3182 -0.188190
## #52 4.13 2.9898 1.140160
## #53 2.75 2.7500 0.000000
## #54 3.25 3.4133 -0.163299
## #55 4.00 3.7966 0.203432
## #56 3.38 3.5909 -0.210874
##
## attr(,"class")
## [1] "qt1"Based on the summary model_qt3 above, color shows the most influential mixture composition that affected ambience of beer.
Besides, we can plot the model to see our new design based on “Bitter” concept
High Grade
- color 6(+), 9(-)
- Illustration 5(+), 3(-)
- Shape 3(+), 2(-)
Ambience
- color 4(+), 10(-)
- Illustration 6(+), 7(-)
- Shape 3(+), 2(-)
Based on the plot above, we can get the conclusion that for the color category, red (Color.6) has the highest positive category score, besides, illustration of a crown(Illustration.5) face indicates a beer that is the most high grade, and for shape category, other of oval or other trad (Shape.3) is the most contribution to high grade.
Based on the plot above, we can get the conclusion that for the color category, blue (Color.4) has the highest positive category score, besides, illustration of other object (Illustration.6) face indicates a beer that is the most ambience, and for shape category, other of oval or other trad (Shape.3) is the most contribution to ambience.
Model linear model
model_lm2 <- lm(High_Grade ~Color+Illustration+Shape, beer)
model_lm3 <- lm(Ambience ~Color+Illustration+Shape, beer)
summary(model_lm2)##
## Call:
## lm(formula = High_Grade ~ Color + Illustration + Shape, data = beer)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.14782 -0.27743 -0.02488 0.22984 1.18940
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.54349 0.39484 6.442 1.41e-07 ***
## Color2 -0.56801 0.34105 -1.665 0.1040
## Color3 0.67192 0.32617 2.060 0.0463 *
## Color4 0.66325 0.35120 1.889 0.0666 .
## Color5 -0.25536 0.35039 -0.729 0.4706
## Color6 0.68941 0.27248 2.530 0.0157 *
## Color7 0.62595 0.43225 1.448 0.1558
## Color8 0.28908 0.40813 0.708 0.4831
## Color9 -0.83583 0.64081 -1.304 0.2000
## Color10 -0.75861 0.70717 -1.073 0.2902
## Illustration2 0.08329 0.48088 0.173 0.8634
## Illustration3 -1.05053 0.73261 -1.434 0.1598
## Illustration4 0.44652 0.40270 1.109 0.2745
## Illustration5 0.70722 0.31431 2.250 0.0303 *
## Illustration6 0.42116 0.42759 0.985 0.3309
## Illustration7 -0.62276 0.45450 -1.370 0.1787
## Shape2 -0.08385 0.54957 -0.153 0.8795
## Shape3 0.21512 0.22749 0.946 0.3503
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.6078 on 38 degrees of freedom
## Multiple R-squared: 0.5829, Adjusted R-squared: 0.3963
## F-statistic: 3.124 on 17 and 38 DF, p-value: 0.00174##
## Call:
## lm(formula = Ambience ~ Color + Illustration + Shape, data = beer)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.06819 -0.28384 -0.02252 0.27519 1.14016
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.915048 0.369566 7.888 1.6e-09 ***
## Color2 -0.343098 0.319225 -1.075 0.2892
## Color3 0.272684 0.305299 0.893 0.3774
## Color4 0.656920 0.328721 1.998 0.0529 .
## Color5 -0.328350 0.327968 -1.001 0.3231
## Color6 0.478378 0.255038 1.876 0.0684 .
## Color7 0.124950 0.404582 0.309 0.7591
## Color8 0.121892 0.382014 0.319 0.7514
## Color9 -0.568190 0.599800 -0.947 0.3495
## Color10 -0.808188 0.661909 -1.221 0.2296
## Illustration2 -0.003268 0.450105 -0.007 0.9942
## Illustration3 -0.330872 0.685722 -0.483 0.6322
## Illustration4 0.433578 0.376924 1.150 0.2572
## Illustration5 0.380002 0.294190 1.292 0.2043
## Illustration6 0.494106 0.400222 1.235 0.2246
## Illustration7 -0.359255 0.425415 -0.844 0.4037
## Shape2 -0.301508 0.514400 -0.586 0.5613
## Shape3 0.023140 0.212932 0.109 0.9140
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5689 on 38 degrees of freedom
## Multiple R-squared: 0.4273, Adjusted R-squared: 0.1712
## F-statistic: 1.668 on 17 and 38 DF, p-value: 0.094165 Conclusion
Note2 : Berarti modelnya udah bagus karena R-squared nya kurang dari 0.6 untuk mengatakaan bahwa ketika can warna merah, ilustratornya crown dan bentuk can nya selain oval atau other trad itu yang menggambarkan high grade.
Note3 : Berarti modelnya udah bagus karena R-squared nya kurang dari 0.6 untuk mengatakaan bahwa ketika can warna biru, tidak ada ilustrasi simbol yang lain atau object yang lain dan bentuk can nya selain oval atau other trad itu yang menggambarkan ambience.
6 Reference
- Lokman, Anitawati Mohd, 2010. Design & Emotion : The Kansei Engineering Methodology. Malaysia: Universiti Teknologi MARA.
- Nagamachi, Mitsuo, 2011. Kansei/Affective Engineering. New York: Taylor and Francis Group.
- Nagamachi, Mitsuo & Lokman, Anitawati Mohd, 2011. Innovations of Kansei Engineering. New York: Taylor and Francis Group.