Tidy Analysis of OSYM Data
Omer Tarkan Acikelim
April 20, 2018
This is an individual in-class exercise. At the end of the exercise, please upload html outputs to your Progress Journals.
- Get the highest
max_scoreprograms from eachexam_type.
load("D:\\Users\\SUUSER\\Desktop\\osym_data_2017_v2.RData")
osym_data_2017 %>% group_by(exam_type) %>% summarise(maxscore = max(max_score))## # A tibble: 19 x 2
## exam_type maxscore
## <chr> <dbl>
## 1 DİL_1 543
## 2 DİL_2 485
## 3 DİL_3 490
## 4 MF 408
## 5 MF_1 537
## 6 MF_2 527
## 7 MF_3 564
## 8 MF_4 575
## 9 TM_1 561
## 10 TM_2 458
## 11 TM_3 558
## 12 TS_1 527
## 13 TS_2 536
## 14 YGS_1 516
## 15 YGS_2 473
## 16 YGS_3 315
## 17 YGS_4 466
## 18 YGS_5 419
## 19 YGS_6 519- Plot the top 10 programs of İSTANBUL ÜNİVERSİTESİ in terms of total quota in a bar chart.
osym_data_2017 <- osym_data_2017 %>% mutate(general_quota= as.numeric(general_quota), general_placement = as.numeric(general_placement))
top_10_iu <- osym_data_2017 %>% filter(university_name == "İSTANBUL ÜNİVERSİTESİ") %>% arrange(desc(general_quota)) %>% slice(1:10)
top_10_iu## # A tibble: 10 x 14
## program_id university_name city faculty_name program_name exam_type
## <chr> <chr> <chr> <chr> <chr> <chr>
## 1 105611289 İSTANBUL ÜNİVER~ İSTAN~ Açık ve Uzak~ Sosyoloji (~ TM_3
## 2 105611298 İSTANBUL ÜNİVER~ İSTAN~ Açık ve Uzak~ İşletme (Aç~ TM_1
## 3 105611183 İSTANBUL ÜNİVER~ İSTAN~ Açık ve Uzak~ Coğrafya (A~ TS_1
## 4 105611305 İSTANBUL ÜNİVER~ İSTAN~ Açık ve Uzak~ Tarih (Açık~ TS_2
## 5 105611314 İSTANBUL ÜNİVER~ İSTAN~ Açık ve Uzak~ İktisat (Aç~ TM_1
## 6 105610501 İSTANBUL ÜNİVER~ İSTAN~ Hukuk Fakült~ Hukuk TM_3
## 7 105611174 İSTANBUL ÜNİVER~ İSTAN~ Açık ve Uzak~ Felsefe (Aç~ TM_3
## 8 105630092 İSTANBUL ÜNİVER~ İSTAN~ Hukuk Fakült~ Hukuk (İÖ) TM_3
## 9 105610634 İSTANBUL ÜNİVER~ İSTAN~ İstanbul Tıp~ İstanbul Tıp MF_3
## 10 105610016 İSTANBUL ÜNİVER~ İSTAN~ Cerrahpaşa T~ Cerrahpaşa ~ MF_3
## # ... with 8 more variables: general_quota <dbl>, general_placement <dbl>,
## # min_score <dbl>, max_score <dbl>, val_quota <dbl>,
## # val_placement <dbl>, val_min_score <dbl>, val_max_score <dbl>library(ggplot2)
ggplot(data=top_10_iu) + geom_bar(aes(x= reorder(program_name, - general_quota),
y=general_quota), stat="identity") + theme(axis.text.x = element_text(angle=90)) 
- Calculate the fill rate (sum(
general_placement)/sum(general_quota)) per city and return the top 10 cities.
osym_data_2017## # A tibble: 11,465 x 14
## program_id university_name city faculty_name program_name exam_type
## <chr> <chr> <chr> <chr> <chr> <chr>
## 1 100110266 ABANT İZZET BAY~ BOLU Bolu Sağlık ~ Hemşirelik YGS_2
## 2 100110487 ABANT İZZET BAY~ BOLU Bolu Turizm ~ Gastronomi v~ YGS_4
## 3 100110724 ABANT İZZET BAY~ BOLU Bolu Turizm ~ Turizm İşlet~ YGS_6
## 4 100130252 ABANT İZZET BAY~ BOLU Bolu Turizm ~ Turizm İşlet~ YGS_6
## 5 100110433 ABANT İZZET BAY~ BOLU Diş Hekimliğ~ Diş Hekimliği MF_3
## 6 100110609 ABANT İZZET BAY~ BOLU Diş Hekimliğ~ Diş Hekimliğ~ MF_3
## 7 100110018 ABANT İZZET BAY~ BOLU Eğitim Fakül~ Bilgisayar v~ MF_1
## 8 100110027 ABANT İZZET BAY~ BOLU Eğitim Fakül~ Fen Bilgisi ~ MF_2
## 9 100110036 ABANT İZZET BAY~ BOLU Eğitim Fakül~ İlköğretim M~ MF_1
## 10 100110045 ABANT İZZET BAY~ BOLU Eğitim Fakül~ İngilizce Öğ~ DİL_1
## # ... with 11,455 more rows, and 8 more variables: general_quota <dbl>,
## # general_placement <dbl>, min_score <dbl>, max_score <dbl>,
## # val_quota <dbl>, val_placement <dbl>, val_min_score <dbl>,
## # val_max_score <dbl>osym_data_2017 %>% group_by(city) %>% summarise(fillrate = sum(general_placement)/sum(general_quota)) %>% arrange(desc(fillrate)) %>% slice(1:10)## # A tibble: 10 x 2
## city fillrate
## <chr> <dbl>
## 1 GEBZE 1.02
## 2 TEKİRDAĞ 1.01
## 3 MANİSA 1.01
## 4 KOCAELİ 1.00
## 5 BURSA 1.00
## 6 SUMGAYIT - AZERBAYCAN 1.00
## 7 DİYARBAKIR 0.999
## 8 DENİZLİ 0.998
## 9 EDİRNE 0.998
## 10 SAKARYA 0.997- Find full (
general_placement==general_quota) Endüstri Mühendisliği programs (usegrepl) and draw a scatterplot of min_score vs max_score. Set transparency parameter (alpha) to 0.7. Set program colors according to whether it is a foundation university or state university. (Tip: State university programs ids start with 1, foundation 2, KKTC 3 and other abroad 4. You can usesubstrfunction.).
library(ggplot2)
fullieprograms <- osym_data_2017%>% select(program_id,university_name,program_name,general_quota,general_placement, min_score, max_score) %>% filter(program_name == "Endüstri Mühendisliği",general_quota==general_placement)
fullieprograms## # A tibble: 15 x 7
## program_id university_name program_name general_quota general_placeme~
## <chr> <chr> <chr> <dbl> <dbl>
## 1 101011032 ANADOLU ÜNİVERS~ Endüstri Mü~ 65.0 65.0
## 2 102910297 ÇUKUROVA ÜNİVER~ Endüstri Mü~ 60.0 60.0
## 3 103110478 DOKUZ EYLÜL ÜNİ~ Endüstri Mü~ 90.0 90.0
## 4 103810198 ESKİŞEHİR OSMAN~ Endüstri Mü~ 80.0 80.0
## 5 104010131 GALATASARAY ÜNİ~ Endüstri Mü~ 25.0 25.0
## 6 104110642 GAZİ ÜNİVERSİTE~ Endüstri Mü~ 90.0 90.0
## 7 105510292 İSTANBUL TEKNİK~ Endüstri Mü~ 80.0 80.0
## 8 105610704 İSTANBUL ÜNİVER~ Endüstri Mü~ 60.0 60.0
## 9 106210829 KARADENİZ TEKNİ~ Endüstri Mü~ 60.0 60.0
## 10 106910337 KOCAELİ ÜNİVERS~ Endüstri Mü~ 80.0 80.0
## 11 102510256 MANİSA CELÂL BA~ Endüstri Mü~ 60.0 60.0
## 12 108210823 ONDOKUZ MAYIS Ü~ Endüstri Mü~ 60.0 60.0
## 13 108810332 SAKARYA ÜNİVERS~ Endüstri Mü~ 100 100
## 14 109710321 ULUDAĞ ÜNİVERSİ~ Endüstri Mü~ 60.0 60.0
## 15 110110225 YILDIZ TEKNİK Ü~ Endüstri Mü~ 80.0 80.0
## # ... with 2 more variables: min_score <dbl>, max_score <dbl>fullieprograms1<- fullieprograms %>% mutate(new_id = substr(program_id,1,1))
ggplot(data=fullieprograms1, aes(x=max_score, y=min_score, color= new_id)) + geom_point(alpha= .7)
- Find the top 10 faculties with the highest quotas and draw a bar chart. Ignore similar names and typos in faculty names.
question5 <- osym_data_2017 %>% arrange(desc(general_quota)) %>% slice(1:10)
ggplot(data=question5) + geom_bar(aes(x=reorder(program_name, -general_quota),y=general_quota), stat="identity") + theme(axis.text.x=element_text(angle=90,size = 8))
- Find all full medicine programs (Tıp but not Tıp Mühendisliği) of foundation universities group by university calculate total quotas per university and maximum max_score and minimum min_score as bounds, ordered and colored by total quota. (Tip: Use
geom_crossbar)
TIP <- osym_data_2017 %>% filter(grepl("Tıp", program_name)) %>% filter(!grepl("Mühendis",program_name)) %>% filter(substr(program_id,1,1) == 1) %>% group_by(university_name) %>% summarize(total_quota = sum(general_quota), maximum_score = max(max_score),minimum_score = min(min_score) ) %>%ungroup() %>% arrange(desc(total_quota))
ggplot(data=TIP) + geom_crossbar(aes(x=reorder(university_name,-total_quota) , y = ((maximum_score+minimum_score)/2) , ymax= maximum_score , ymin = minimum_score , color = total_quota)) + labs (y = "Max and Min Scores") + theme(axis.text.x=element_text(angle=90,hjust=0.5,vjust=0.5)) + xlab ("University")
- Freestyle: Compare Mechanical Engineering and Civil Engineering departments and have an idea about student’s demand to these departments. (while max score for civil engineering departments is higher than mechanical engineering departmens in average, fill rate to the civil engineering departments is still higher than to the mechanical engineering programs)
civil <-osym_data_2017 %>% select (university_name, program_name, max_score, general_placement, general_quota ) %>% filter(program_name == "İnşaat Mühendisliği" ) %>% summarise(mean_civil = mean(max_score) , fillrate = sum(general_placement)/sum(general_quota))
civil## # A tibble: 1 x 2
## mean_civil fillrate
## <dbl> <dbl>
## 1 375 1.01mechanical <-osym_data_2017 %>% select (university_name, program_name, max_score, general_placement, general_quota ) %>% filter(program_name == "Makine Mühendisliği" ) %>% summarise(mean_civil = mean(max_score) , fillrate = sum(general_placement)/sum(general_quota))
mechanical## # A tibble: 1 x 2
## mean_civil fillrate
## <dbl> <dbl>
## 1 341 0.967Result : While max score for civil engineering departments is higher than mechanical engineering departmens in average, fill rate to the civil engineering departments is still higher than to the mechanical engineering programs.
- Freestyle: Compare Koç University with Bilkent University in terms of their IE Departments
bilkentvskoc <- osym_data_2017 %>% select(university_name , program_name, faculty_name , max_score) %>% filter( faculty_name == "Mühendislik Fakültesi",program_name =="Endüstri Mühendisliği (İngilizce) (Tam Burslu)" | program_name == "Endüstri Mühendisliği (İngilizce) (%50 Burslu)" | program_name == "Endüstri Mühendisliği (İngilizce) (%25 Burslu)",university_name =="İHSAN DOĞRAMACI BİLKENT ÜNİVERSİTESİ"| university_name == "KOÇ ÜNİVERSİTESİ")
bilkentvskoc## # A tibble: 5 x 4
## university_name program_name faculty_name max_score
## <chr> <chr> <chr> <dbl>
## 1 İHSAN DOĞRAMACI BİLKENT ÜNİVERSİTESİ Endüstri Mü~ Mühendislik~ 486
## 2 İHSAN DOĞRAMACI BİLKENT ÜNİVERSİTESİ Endüstri Mü~ Mühendislik~ 515
## 3 KOÇ ÜNİVERSİTESİ Endüstri Mü~ Mühendislik~ 497
## 4 KOÇ ÜNİVERSİTESİ Endüstri Mü~ Mühendislik~ 507
## 5 KOÇ ÜNİVERSİTESİ Endüstri Mü~ Mühendislik~ 540ggplot(data=bilkentvskoc) + geom_bar(aes(x=program_name, y= max_score, color= university_name), stat ="identity")