3.1.1 인기있는 프로젝트 유형

이 질문은 프로젝트 스타터의 관점에서 카테고리 (데이터 집합의 main_category라고 함)와 하위 카테고리 (데이터 집합의 카테고리라고 함)의 두 가지 수준을 기준으로 답한다. 먼저 카테고리별로 프로젝트 수를 조사해보겠다.

cat.freq <- ksdata %>%
  group_by(main_category) %>%
  summarize(count=n()) %>%
  arrange(desc(count))

cat.freq$main_category <- factor(cat.freq$main_category, levels=cat.freq$main_category)

ggplot(cat.freq, aes(main_category, count, fill=count)) + geom_bar(stat="identity") + 
  ggtitle("Projects by Category") + xlab("Project Category") + ylab("Frequency") + 
  geom_text(aes(label=count), vjust=-0.5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12, angle=90), legend.position="null") + 
  scale_fill_gradient(low="skyblue1", high="royalblue4")

3.1.2 인기있는 하위 카테고리 프로젝트 유형

subcat.freq <- ksdata %>%
  group_by(category) %>%
  summarize(count=n()) %>%
  arrange(desc(count))

subcat.freq$category <- factor(subcat.freq$category, levels=subcat.freq$category)

ggplot(head(subcat.freq, 10), aes(category, count, fill=count)) + geom_bar(stat="identity") + 
  ggtitle("Projects by Subcategory") + xlab("Project Subcategory") + ylab("Frequency") + 
  geom_text(aes(label=count), vjust=-0.5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12, angle=90), legend.position="null") + 
  scale_fill_gradient(low="skyblue1", high="royalblue4")

3.1.3 카테고리별 추이

# Converting the launch and deadline Dates to correct format:
ksdata$launch_date <- as.Date(ksdata$launched, "%Y-%m-%d")
ksdata$deadline_date <- as.Date(ksdata$deadline, "%Y-%m-%d")

# also addding month and year columns for the deadline and launch dates:
ksdata$launch_year <- substr(ksdata$launched, 1,4)
ksdata$launch_mth <- substr(ksdata$launch_date, 1,7)

ksdata$final_year <- substr(ksdata$deadline,1,4)
ksdata$final_mth <- substr(ksdata$deadline, 1, 7)


ggplot(data=ksdata, aes(x=launch_year)) +
  geom_bar(colour="black", fill=fillColor2) +
  ylab('Count') +
  facet_wrap(~main_category) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
  theme_minimal()

3.2.1 모금이 가장 많이 된 프로젝트

위의 질문과 비슷하게 느껴질 수 있지만, 후원자의 관점에서 가장 인기 있는 프로젝트를 금액 기준으로 정렬해보았다(상위 15개).

kable(head(ksdata[order(-ksdata$usd_pledged), c(2,3,13)], 15))

kable(head(ksdata[order(-ksdata$backers), c(2,3,11)], 15))

3.2.2 카테고리별 모금 총액

pledged.tot <- ksdata %>%
  group_by(main_category) %>%
  summarize(total=sum(usd_pledged)) %>%
  arrange(desc(total))

pledged.tot$main_category <- factor(pledged.tot$main_category, levels=pledged.tot$main_category)

ggplot(pledged.tot, aes(main_category, total/1000000, fill=total)) + geom_bar(stat="identity") + 
  ggtitle("Total Amount Pledged by Category") + xlab("Project Category") + 
  ylab("Amount Pledged (USD millions)") + 
  geom_text(aes(label=paste0("$", round(total/1000000,1))), vjust=-0.5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12, angle=90), legend.position="null") + 
  scale_fill_gradient(low="skyblue1", high="royalblue4")

3.2.3 1인당 평균 후원금액

pledged.avg <- ksdata %>%
  group_by(main_category) %>%
  summarize(pledged=sum(usd_pledged), backers=sum(backers)) %>%
  mutate(avg=pledged/backers) %>%
  arrange(desc(avg))

pledged.avg$main_category <- factor(pledged.avg$main_category, levels=pledged.avg$main_category)

ggplot(pledged.avg, aes(main_category, avg, fill=avg)) + geom_bar(stat="identity") + 
  ggtitle("Average Amount Pledged per Backer") + xlab("Project Category") + 
  ylab("Amount Pledged (USD)") + 
  geom_text(aes(label=paste0("$", round(avg,2))), vjust=-0.5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12, angle=90), legend.position="null") + 
  scale_fill_gradient(low="skyblue1", high="royalblue4")

3.2.4 Box plot으로 데이터의 분포 확인하기

ggplot(ksdata, aes(main_category, usd_pledged, fill=main_category)) + geom_boxplot() + 
  ggtitle("Amount Pledged vs. Project Category") + xlab("Project Category") + 
  ylab("Amount Pledged (USD)") + 
  theme(plot.title=element_text(size=15, face="bold", hjust=0.5), 
        axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12, angle=90), legend.position="null") + 
  coord_cartesian(ylim=c(0,20000))

3.3.1 목표 금액이 가장 높은 프로젝트

자금 목표가 가장 높은 상위 15개 프로젝트를 살펴보겠다.

kable(head(ksdata[order(-ksdata$usd_goal), c(2,3,14,10)], 15))

3.3.2 성공한 프로젝트 중 목표 금액이 가장 높은 프로젝트

goal.tops <- ksdata[ksdata$state=="successful",]

kable(head(goal.tops[order(-goal.tops$usd_goal), c(2,3,14,10)], 15))

3.3.3 각 카테고리의 평균 프로젝트 목표 금액

goal.avg <- ksdata %>%
  group_by(main_category) %>%
  summarize(goal=sum(usd_goal), projects=n()) %>%
  mutate(avg=goal/projects) %>%
  arrange(desc(avg))

goal.avg$main_category <- factor(goal.avg$main_category, levels=goal.avg$main_category)

ggplot(goal.avg, aes(main_category, avg, fill=avg)) + geom_bar(stat="identity") + 
  ggtitle("Average Project Goal") + xlab("Project Category") + ylab("Project Goal (USD)") + 
  geom_text(aes(label=paste0("$", round(avg,0))), vjust=-0.5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12, angle=90), legend.position="null") + 
  scale_fill_gradient(low="skyblue1", high="royalblue4")

3.3.4 프로젝트 카테고리별 목표 금액

ggplot(ksdata, aes(main_category, usd_goal, fill=main_category)) + geom_boxplot() + 
  ggtitle("Project Goal vs. Project Category") + xlab("Project Category") +
  ylab("Project Goal (USD)") + 
  theme(plot.title=element_text(size=15, face="bold", hjust=0.5), 
        axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12, angle=90), legend.position="null") + 
  coord_cartesian(ylim=c(0,60000))

3.3.5 목표 금액과 모금액의 분포

usd.amounts <- gather(ksdata, type, amount, usd_pledged, usd_goal, factor_key=T)

ggplot(usd.amounts, aes(amount, fill=type)) + 
  geom_histogram(alpha=0.5, position="identity") + 
  ggtitle("Distribution of USD Pledged vs. USD Goal") + xlab("USD") + 
  ylab("Frequency") + scale_fill_discrete("Type", labels=c("USD Pledged", "USD Goal"))

usd.amounts <- gather(ksdata, type, amount, usd_pledged, usd_goal, factor_key=T)

ggplot(usd.amounts, aes(log(amount+1), fill=type)) + 
  geom_histogram(alpha=0.5, position="identity") + 
  ggtitle("Distribution of log(USD Pledged) vs. log(USD Goal)") + xlab("log(USD + 1)") + 
  ylab("Frequency") + scale_fill_discrete("Type", labels=c("USD Pledged", "USD Goal"))

3.3.6 목표 달성률이 가장 높았던 프로젝트

ksdata$ratio <- ksdata$usd_pledged/ksdata$usd_goal

kable(head(ksdata[order(-ksdata$ratio), c(2,3,13,14,15)], 15))

3.3.7 목표 달성률이 가장 높았던 프로젝트 (목표 금액 $1,000 이상)

goal.min <- ksdata[ksdata$usd_goal>=1000,]

kable(head(goal.min[order(-goal.min$ratio), c(2,3,13,14,15)], 15))

3.4.1 프로젝트 상태별 분석

프로젝트 상태(성공, 실패, 취소 등)별로 세분화해서 분석해보겠다.

state.freq <- ksdata %>%
  group_by(state) %>%
  summarize(count=n()) %>%
  arrange(desc(count))

state.freq$state <- factor(state.freq$state, levels=state.freq$state)

ggplot(state.freq, aes(state, count, fill=count)) + geom_bar(stat="identity") + 
  ggtitle("Projects by Status") + xlab("Project Status") + ylab("Frequency") + 
  geom_text(aes(label=count), vjust=-0.5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12), legend.position="null") + 
  scale_fill_gradient(low="skyblue1", high="royalblue4")

3.4.2 좀 더 세분화(그룹화)된 프로젝트 상태별 분석

state.grp <- ksdata %>%
  filter(state!="undefined") %>%
  mutate(grp=ifelse(state %in% c("successful", "failed"), "complete", "incomplete")) %>%
  group_by(grp, state) %>%
  summarize(count=n()) %>%
  mutate(pct=count/sum(count)) %>%
  arrange(grp, desc(state))

state.grp$state <- factor(state.grp$state, levels=state.grp$state)

ggplot(state.grp, aes(grp, pct, fill=state)) + geom_bar(stat="identity") + 
  ggtitle("Project Status by Completion") + xlab("Project Completion") + ylab("Percentage") + 
  scale_x_discrete(labels=c("Complete", "Incomplete")) + 
  scale_y_continuous(labels=scales::percent) + 
  scale_fill_brewer(name="Project Status", 
                    labels=c("Successful", "Failed", "Suspended", "Live", "Cancelled"), 
                    palette="Set1") + 
  geom_text(aes(label=paste0(round(pct*100,1),"%")), position=position_stack(vjust=0.5), 
            colour="white", size=5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12), legend.position="bottom", 
        legend.title=element_text(size=12, face="bold"))

3.4.3 카테고리별 프로젝트 성공/실패 여부

state.pct <- ksdata %>%
  filter(state %in% c("successful", "failed")) %>%
  group_by(main_category, state) %>%
  summarize(count=n()) %>%
  mutate(pct=count/sum(count)) %>%
  arrange(desc(state), pct)

state.pct$main_category <- factor(state.pct$main_category, 
                                  levels=state.pct$main_category[1:(nrow(state.pct)/2)])

ggplot(state.pct, aes(main_category, pct, fill=state)) + geom_bar(stat="identity") + 
  ggtitle("Success vs. Failure Rate by Project Category") + 
  xlab("Project Category") + ylab("Percentage") + scale_y_continuous(labels=scales::percent) + 
  scale_fill_discrete(name="Project Status", breaks=c("successful", "failed"),
                      labels=c("Success", "Failure")) + 
  geom_text(aes(label=paste0(round(pct*100,1),"%")), position=position_stack(vjust=0.5), 
            colour="white", size=5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12), legend.position="bottom", 
        legend.title=element_text(size=12, face="bold")) + coord_flip()

3.5.1 프로젝트 기간별 성공률

ksdata$length <- interval(ymd_hms(ksdata$launched), ymd(ksdata$deadline)) %/% days(1)

length.pct <- ksdata %>%
  filter(state %in% c("successful", "failed"), length <= 61) %>%
  group_by(length, state) %>%
  summarize(count=n()) %>%
  mutate(pct=count/sum(count))

ggplot(length.pct[length.pct$state=="successful",], aes(length, pct)) + 
  geom_point(colour="royalblue4", size=2.5) + ggtitle("Success Rate vs. Project Length") + 
  xlab("Project Length (Days)") + ylab("Success Rate (%)") + 
  scale_x_continuous(breaks=c(0,10,20,30,40,50,60)) + geom_vline(xintercept=30, colour="red") + 
  theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"))

3.5.2 프로젝트 기간에 따른 프로젝트 분포

ggplot(ksdata[ksdata$length <= 61,], aes(length)) + geom_density(colour="royalblue4", size=1) + 
  ggtitle("Distribution of Projects by Campaign Length") + xlab("Project Length (Days)") + 
  ylab("Density (%)") + scale_x_continuous(breaks=c(0,10,20,30,40,50,60)) + 
  geom_vline(xintercept=30, colour="red") + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"))

3.6.1 연도별 프로젝트 런칭 수

year.freq <- ksdata %>%
  filter(year(launched)!="1970") %>%
  group_by(year=year(launched)) %>%
  summarize(count=n())

ggplot(year.freq, aes(year, count, fill=count)) + geom_bar(stat="identity") + 
  ggtitle("Number of Projects by Launch Year") + xlab("Year") + ylab("Frequency") + 
  scale_x_discrete(limits=c(2009:2018)) + 
  geom_text(aes(label=paste0(count)), vjust=-0.5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12), legend.position="null") + 
  scale_fill_gradient(low="skyblue1", high="royalblue4")

3.6.2 연도별 프로젝트 성공/실패율

state.pct2 <- ksdata %>%
  filter(year(launched)!="1970", state %in% c("successful", "failed")) %>%
  group_by(year=year(launched), state) %>%
  summarize(count=n()) %>%
  mutate(pct=count/sum(count)) %>%
  arrange(desc(state))

ggplot(state.pct2, aes(year, pct, fill=state)) + geom_bar(stat="identity") + 
  ggtitle("Success vs. Failure Rate by Year Launched") + 
  xlab("Year") + ylab("Percentage") + scale_x_discrete(limits=c(2009:2017)) + 
  scale_y_continuous(labels=scales::percent) + 
  scale_fill_discrete(name="Project Status", breaks=c("successful", "failed"),
                      labels=c("Success", "Failure")) + 
  geom_text(aes(label=paste0(round(pct*100,1),"%")), position=position_stack(vjust=0.5), 
            colour="white", size=5) + theme_economist() + 
  theme(plot.title=element_text(hjust=0.5), axis.title=element_text(size=12, face="bold"), 
        axis.text.x=element_text(size=12), legend.position="bottom", 
        legend.title=element_text(size=12, face="bold"))

3.6.3 년도/프로젝트 유형별 차이 비교 시각화

cat.year <- ksdata %>%
  filter(!year(launched) %in% c("1970", "2018")) %>%
  group_by(main_category, year=year(launched)) %>%
  summarize(count=n())

cat.year2 <- t(matrix(cat.year$count, nrow=9))
colnames(cat.year2) <- c("2009", "2010", "2011", "2012", "2013", "2014", "2015", "2016", "2017")
rownames(cat.year2) <- levels(ksdata$main_category)

heatmap.2(cat.year2, dendrogram="row", Colv=F, trace="none", margins=c(10,10))

3.7.1 국가별 프로젝트 성공 현황

ks5 <- ksdata %>%
  group_by(country, state) %>%
  select(goal, pledged, backers) %>%
  summarise(freq=n()) %>%
  mutate(categoryPerc = freq/sum(freq)) %>%
  filter(state == "successful") %>%
  arrange(desc(categoryPerc))

ks55 <- head(ks5, 10)

p5 <- ggplot(ks55, aes(x = reorder(country, -categoryPerc), y = round(categoryPerc*100, digits = 1))) +
         geom_bar(stat = "identity", fill = fillColor2) + theme(axis.text.x = element_text(angle=60, hjust=1)) + geom_text(aes(label=round(categoryPerc*100, digits = 1)), position=position_dodge(width=0.9), vjust=-0.25) +
  theme_economist()

p5+labs(y="Success %", x="Country", title="Top 10 Country by Successful project %")

3.7.2 국가별 프로젝트 모금 현황

ks6 <- ksdata %>%
  group_by(country, state) %>%
  select(goal, pledged, backers) %>%
  summarise(sumpledge=sum(pledged)) %>%
  filter(state == "successful") %>%
  arrange(desc(sumpledge))
 
ks66 <- head(ks6, 10)
 
p6 <- ggplot(ks66, aes(x = reorder(country, -sumpledge), y = round(sumpledge/1000000, digits = 0))) +
         geom_bar(stat = "identity", fill = fillColor2) + theme(axis.text.x = element_text(angle=60, hjust=1)) + geom_text(aes(label=round(sumpledge/1000000, digits = 0)), position=position_dodge(width=0.9), vjust=-0.25) +
  theme_economist()

p6 + labs(y="Money raised (USD Million)", x="Country", title="Top 10 Country by Money raised (in USD Million)")

3.7.3 미국과 미국이 아닌 국가들의 프로젝트 성공률 비교

킥스타터는 전세계적으로 사용되고 있기 때문에, 카테고리 성공 여부는 프로젝트가 미국 또는 미국 이외의 지역에서 시작되었는지에 따라 결정된다. 이를 위해 데이터를 정리하고 분류하여 미국과 미국이 아닌 지역에서 성공률로 나누었다(아래의 차트는 미국 및 미국 외 국가의 카테고리 성공율을 비교한 것이다). 전반적으로 미국의 성공률이 조금 더 높다.

ks7 <- ksdata %>%
  filter(country %in% c('US')) %>%
  group_by(main_category, state) %>%
  select(goal, pledged, backers) %>%
  summarise(freq=n()) %>%
  mutate(us = freq/sum(freq)) %>%
  filter(state == "successful") %>%
  arrange(desc(us))

us_success <- select(ks7, main_category, us)


ks8 <- ksdata %>%
  filter(!country %in% c('US')) %>%
  group_by(main_category, state) %>%
  select(goal, pledged, backers) %>%
  summarise(freq=n()) %>%
  mutate(non_us = freq/sum(freq)) %>%
  filter(state == "successful") %>%
  arrange(desc(non_us))

non_us_success <- select(ks8, main_category, non_us)


p7 <- ggplot(ks7, aes(x = reorder(main_category, -us), y = round(us*100, digits = 0))) +
         geom_bar(stat = "identity", fill = fillColor2) + theme(axis.text.x = element_text(angle=60, hjust=1)) + geom_text(aes(label=round(us*100, digits = 0)), position=position_dodge(width=0.3), vjust=-0.25) + labs(y="Success %", x="Main Category", title="USA - Successful project %") + theme_economist()+ theme(axis.text.x = element_text(angle = 90, hjust = 1))



p8 <- ggplot(ks8, aes(x = reorder(main_category, -non_us), y = round(non_us*100, digits = 0))) +
         geom_bar(stat = "identity", fill = fillColor2) + theme(axis.text.x = element_text(angle=60, hjust=1)) + geom_text(aes(label=round(non_us*100, digits = 0)), position=position_dodge(width=0.3), vjust=-0.25) + labs(y="Success %", x="Main Category", title="Non USA - Successful project %") + theme_economist() + theme(axis.text.x = element_text(angle = 90, hjust = 1))


ggarrange(p7, p8, ncol = 2)

4.1.1 독립성 검정 (피어슨 카이제곱 검정)

\[H0(귀무가설)\] 두 개의 범주형 변수 간에는 연관성이 없고, 행 변수와 열 변수는 모집단에서 독립적이며, 카테고리의 성공률에 영향을 미친다. \[H1(대립가설)\] 두 개의 범주형 변수 간에는 연관성이 없고, 행 변수와 열 변수는 독립적이지 않으며, 카테고리의 성공에 영향을 주지 않는다.

compare_main <- left_join(us_success, non_us_success, by = c("main_category" = "main_category"))
compare_main

## # A tibble: 15 x 3
## # Groups:   main_category [?]
##    main_category    us non_us
##    <fct>         <dbl>  <dbl>
##  1 Dance         0.647  0.463
##  2 Theater       0.599  0.596
##  3 Comics        0.541  0.535
##  4 Music         0.502  0.286
##  5 Art           0.426  0.345
##  6 Film & Video  0.381  0.329
##  7 Games         0.380  0.298
##  8 Design        0.354  0.342
##  9 Photography   0.318  0.274
## 10 Publishing    0.314  0.287
## 11 Food          0.268  0.158
## 12 Fashion       0.248  0.236
## 13 Crafts        0.247  0.220
## 14 Journalism    0.225  0.176
## 15 Technology    0.219  0.155

chisq.test(data.frame(compare_main$us, compare_main$non_us))

## 
##  Pearson's Chi-squared test
## 
## data:  data.frame(compare_main$us, compare_main$non_us)
## X-squared = 0.074268, df = 14, p-value = 1

4.1.2 분산 분석 (Anova Test)

킥스타터에서 ‘Games’와 ’Design’ 카테고리는 가장 많은 모금액($741, $734 millon)이 형성되었는데, 이 카테고리에 의해 모금된 금액이 통계적으로 차이가 있는지 아니면 비슷한지 확인해보겠다. 분석을 위해 상위 1,000개의 프로젝트를 선택하고 Anova 분석을 시행하겠다.

ks9 <- ksdata %>%
  filter(main_category %in% c('Games')) %>%
  select(pledged) %>%
  arrange(desc(pledged))

ks10 <- ksdata %>%
  filter(main_category %in% c('Design')) %>%
  select(pledged) %>%
  arrange(desc(pledged))

ks9 <- head(ks9, 1000)
ks10 <- head(ks10, 1000)

sprintf("Mean amount pledged on Design Category: %f", mean(ks9$pledged))

## [1] "Mean amount pledged on Design Category: 500963.445080"

## [1] "Mean amount pledged on Design Category: 399344.450060"
sprintf("Mean amount pledged on Games Category: %f", mean(ks10$pledged))

## [1] "Mean amount pledged on Games Category: 519495.490640"

## [1] "Mean amount pledged on Games Category: 387663.683480"
#ks11 <- round(ks9[sample(1:nrow(ks9), 1000, replace = FALSE), ], digits = -3)
#ks12 <- round(ks10[sample(1:nrow(ks10), 1000, replace = FALSE), ], digits = -3)
ks9 <- round(ks9, digits= -3)
ks10 <- round(ks10, digits = -3)
scores = data.frame(ks9, ks10)
#boxplot(scores)
scores = stack(scores)
resultaov <- aov(values ~ ind, data=scores, var.equal = T)
summary(resultaov)

##               Df    Sum Sq   Mean Sq F value Pr(>F)
## ind            1 1.718e+11 1.718e+11   0.165  0.684
## Residuals   1998 2.077e+15 1.040e+12

4.1.3 국가별 차이 비교 시각화

countries.freq <- ksdata %>%
  filter(country!='N,0"') %>%
  group_by(country) %>%
  summarize(count=n())

countries.match <- joinCountryData2Map(countries.freq, joinCode="ISO2", nameJoinColumn="country")

## 22 codes from your data successfully matched countries in the map
## 0 codes from your data failed to match with a country code in the map
## 220 codes from the map weren't represented in your data

mapCountryData(countries.match, nameColumnToPlot="count", 
               mapTitle="Number of Projects by Country", catMethod="logFixedWidth", 
               colourPalette="heat")

4.2.1 Linear Regression Model

\[Pledgedamount = A_{0} + B_{1} * backers\]

xyplot(ksdata$pledged~ksdata$backers, data = ksdata, xlab= "Number of backers", ylab= "Amount pledged (USD)", panel = function(x, y, ...) {
                 panel.xyplot(x, y, ...)
                 panel.lmline(x, y, ...)
               })

4.2.2 선형 회귀 모델의 가정

회귀 분석의 결과와 회귀 분석의 가정을 만족시키는 것을 보여주는 차트는 아래에 표시해두었다.

resultslm <- lm(ksdata$pledged~ksdata$backers, data = ksdata)
summary(resultslm)

## 
## Call:
## lm(formula = ksdata$pledged ~ ksdata$backers, data = ksdata)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -7803302    -2280    -1773    -1674 14405275 
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)    1698.8299   109.0555   15.58   <2e-16 ***
## ksdata$backers   75.5950     0.1194  633.09   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 66660 on 378659 degrees of freedom
## Multiple R-squared:  0.5142, Adjusted R-squared:  0.5142 
## F-statistic: 4.008e+05 on 1 and 378659 DF,  p-value: < 2.2e-16

par(mfrow=c(2,2))
plot(resultslm)