Web Analytics Made Easy - Statcounter

EDA on islam religion growth worldwide

EDA
R
Exploring World Religion dataset to uncover the growth of Islam in various countries using R
Author

invictus

Published

August 14, 2024

Modified

August 20, 2024

Setup

library(tidyverse)
library(visdat)
library(DataExplorer)
library(naniar)
library(janitor)
library(validate)
library(countrycode)
df <- read_csv('./ThrowbackDataThursday 201912 - Religion.csv')
Rows: 2308 Columns: 76
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr  (1): StateNme
dbl (75): index, Year, Christianity - Protestants (Population), Christianity...

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Inspect Dataset

Understand the data

df
# A tibble: 2,308 × 76
   index  Year `Christianity - Protestants (Population)` Christianity - Roman …¹
   <dbl> <dbl>                                     <dbl>                   <dbl>
 1     0  1945                                  66069671                38716742
 2     1  1950                                  73090083                42635882
 3     2  1955                                  79294628                46402368
 4     3  1960                                  90692928                50587880
 5     4  1965                                  94165803                64761783
 6     5  1970                                  93918347                69119143
 7     6  1975                                  93578909                77181399
 8     7  1980                                 101216516                75149580
 9     8  1985                                 107309700                82463100
10     9  1990                                 132040181                65530789
# ℹ 2,298 more rows
# ℹ abbreviated name: ¹​`Christianity - Roman Catholics (Population)`
# ℹ 72 more variables: `Christianity - Eastern Orthodox (Population)` <dbl>,
#   `Christianity - Anglican (Population)` <dbl>,
#   `Christianity - Others (Population)` <dbl>,
#   `Christianity - Total (Population)` <dbl>,
#   `Judaism - Orthodox (Population)` <dbl>, …
df |> names()
 [1] "index"                                       
 [2] "Year"                                        
 [3] "Christianity - Protestants (Population)"     
 [4] "Christianity - Roman Catholics (Population)" 
 [5] "Christianity - Eastern Orthodox (Population)"
 [6] "Christianity - Anglican (Population)"        
 [7] "Christianity - Others (Population)"          
 [8] "Christianity - Total (Population)"           
 [9] "Judaism - Orthodox (Population)"             
[10] "Judaism - Conservatives (Population)"        
[11] "Judaism - Reform (Population)"               
[12] "Judaism - Others (Population)"               
[13] "Judaism - Total (Population)"                
[14] "Islam - Sunni (Population)"                  
[15] "Islam - Shi'a (Population)"                  
[16] "Islam - Ibadhi (Population)"                 
[17] "Islam - Nation of Islam (Population)"        
[18] "Islam - Alawite (Population)"                
[19] "Islam - Ahmadiyya (Population)"              
[20] "Islam - Others (Population)"                 
[21] "Islam - Total (Population)"                  
[22] "Buddhism - Mahayana (Population)"            
[23] "Buddhism - Theravada (Population)"           
[24] "Buddhism - Others (Population)"              
[25] "Buddhism - Total (Population)"               
[26] "Zoroastrian - Total (Population)"            
[27] "Hindu - Total (Population)"                  
[28] "Sike - Total (Population)"                   
[29] "Shinto - Total (Population)"                 
[30] "Baha'i - Total (Population)"                 
[31] "Taoism - Total (Population)"                 
[32] "Confucianism - Total (Population)"           
[33] "Jain - Total (Population)"                   
[34] "Syncretic Religions - Total (Population)"    
[35] "Animist - Total (Population)"                
[36] "Non Religious - Total (Population)"          
[37] "Other Religions - Total (Population)"        
[38] "Total Religious - Total (Population)"        
[39] "Population - Total (Population)"             
[40] "Christianity - Protestants (Percent)"        
[41] "Christianity - Roman Catholics (Percent)"    
[42] "Christianity - Eastern Orthodox (Percent)"   
[43] "Christianity - Anglican (Percent)"           
[44] "Christianity - Others (Percent)"             
[45] "Christianity - Total (Percent)"              
[46] "Judaism - Orthodox (Percent)"                
[47] "Judaism - Conservatives (Percent)"           
[48] "Judaism - Reform (Percent)"                  
[49] "Judaism - Others (Percent)"                  
[50] "Judaism - Total (Percent)"                   
[51] "Islam - Sunni (Percent)"                     
[52] "Islam - Shi'a (Percent)"                     
[53] "Islam - Ibadhi (Percent)"                    
[54] "Islam - Nation of Islam (Percent)"           
[55] "Islam - Alawite (Percent)"                   
[56] "Islam - Ahmadiyya (Percent)"                 
[57] "Islam - Others (Percent)"                    
[58] "Islam - Total (Percent)"                     
[59] "Buddhism - Mahayana (Percent)"               
[60] "Buddhism - Theravada (Percent)"              
[61] "Buddhism - Others (Percent)"                 
[62] "Buddhism - Total (Percent)"                  
[63] "Zoroastrian - Total (Percent)"               
[64] "Hindu - Total (Percent)"                     
[65] "Sike - Total (Percent)"                      
[66] "Shinto - Total (Percent)"                    
[67] "Baha'i - Total (Percent)"                    
[68] "Taoism - Total (Percent)"                    
[69] "Confucianism - Total (Percent)"              
[70] "Jain - Total (Percent)"                      
[71] "Syncretic Religions - Total (Percent)"       
[72] "Animist - Total (Percent)"                   
[73] "Non Religious - Total (Percent)"             
[74] "Other Religions - Total (Percent)"           
[75] "Total Religious - Total (Percent)"           
[76] "StateNme"

Check Data Quality

Missing Values

df |> 
  vis_miss()

df |> 
  vis_expect(function(x) x != 0)

df |> 
  replace_with_na_all(~ . == 0) |> 
  vis_miss()

Total Population Columns
df |> 
  select(contains('Total') & contains('Population')) |> 
  rename_with(function(x) str_remove(x, ' - Total \\(Population\\)')) |> 
  replace_with_na_all(~ . == 0) |> 
  vis_miss(sort_miss = TRUE)

Total Percent Columns
df |> 
  select(contains('Total') & contains('Percent')) |> 
  rename_with(function(x) str_remove(x, ' - Total \\(Percent\\)')) |> 
  replace_with_na_all(~ . == 0) |> 
  vis_miss(sort_miss = TRUE)

Population but Non Total columns
df |> 
  select(-contains('Total') & contains('Population')) |>
  rename_with(function(x) str_remove(x, ' \\(Population\\)')) |> 
  replace_with_na_all(~ . == 0) |> 
  vis_miss(sort_miss = TRUE)

Clean Data

Pick relevant columns

df_islam <- df |> 
  select(Country = StateNme, Year, starts_with('Islam'))

df_islam
# A tibble: 2,308 × 18
   Country                   Year Islam - Sunni (Popula…¹ Islam - Shi'a (Popul…²
   <chr>                    <dbl>                   <dbl>                  <dbl>
 1 United States of America  1945                       0                      0
 2 United States of America  1950                       0                      0
 3 United States of America  1955                       0                      0
 4 United States of America  1960                       0                      0
 5 United States of America  1965                       0                      0
 6 United States of America  1970                       0                      0
 7 United States of America  1975                       0                      0
 8 United States of America  1980                 2407464                 254121
 9 United States of America  1985                 2186675                 298183
10 United States of America  1990                 2019386                 275371
# ℹ 2,298 more rows
# ℹ abbreviated names: ¹​`Islam - Sunni (Population)`,
#   ²​`Islam - Shi'a (Population)`
# ℹ 14 more variables: `Islam - Ibadhi (Population)` <dbl>,
#   `Islam - Nation of Islam (Population)` <dbl>,
#   `Islam - Alawite (Population)` <dbl>,
#   `Islam - Ahmadiyya (Population)` <dbl>, …

Check Data type

df_islam |> 
  str()
tibble [2,308 × 18] (S3: tbl_df/tbl/data.frame)
 $ Country                             : chr [1:2308] "United States of America" "United States of America" "United States of America" "United States of America" ...
 $ Year                                : num [1:2308] 1945 1950 1955 1960 1965 ...
 $ Islam - Sunni (Population)          : num [1:2308] 0 0 0 0 0 ...
 $ Islam - Shi'a (Population)          : num [1:2308] 0 0 0 0 0 ...
 $ Islam - Ibadhi (Population)         : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Nation of Islam (Population): num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Alawite (Population)        : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Ahmadiyya (Population)      : num [1:2308] 0 0 0 0 0 ...
 $ Islam - Others (Population)         : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Total (Population)          : num [1:2308] 0 0 0 0 0 ...
 $ Islam - Sunni (Percent)             : num [1:2308] 0 0 0 0 0 0 0 0.0106 0.0092 0.0081 ...
 $ Islam - Shi'a (Percent)             : num [1:2308] 0 0 0 0 0 0 0 0.0011 0.0013 0.0011 ...
 $ Islam - Ibadhi (Percent)            : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Nation of Islam (Percent)   : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Alawite (Percent)           : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Ahmadiyya (Percent)         : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Others (Percent)            : num [1:2308] 0 0 0 0 0 0 0 0 0 0 ...
 $ Islam - Total (Percent)             : num [1:2308] 0 0 0 0 0 0 0 0.0117 0.0104 0.0092 ...

Fix Data Types

df_islam_correct_type <- df_islam |> 
  mutate(Country = as_factor(Country))

Check Duplicate

df_islam_correct_type |> names()
 [1] "Country"                             
 [2] "Year"                                
 [3] "Islam - Sunni (Population)"          
 [4] "Islam - Shi'a (Population)"          
 [5] "Islam - Ibadhi (Population)"         
 [6] "Islam - Nation of Islam (Population)"
 [7] "Islam - Alawite (Population)"        
 [8] "Islam - Ahmadiyya (Population)"      
 [9] "Islam - Others (Population)"         
[10] "Islam - Total (Population)"          
[11] "Islam - Sunni (Percent)"             
[12] "Islam - Shi'a (Percent)"             
[13] "Islam - Ibadhi (Percent)"            
[14] "Islam - Nation of Islam (Percent)"   
[15] "Islam - Alawite (Percent)"           
[16] "Islam - Ahmadiyya (Percent)"         
[17] "Islam - Others (Percent)"            
[18] "Islam - Total (Percent)"

Check duplicate Year

df_islam_correct_type |> 
  get_dupes() |> 
  relocate(dupe_count)
No variable names specified - using all columns.
# A tibble: 626 × 19
   dupe_count Country  Year `Islam - Sunni (Population)` Islam - Shi'a (Popula…¹
        <int> <fct>   <dbl>                        <dbl>                   <dbl>
 1          2 Cuba     1945                            0                       0
 2          2 Cuba     1945                            0                       0
 3          2 Cuba     1950                            0                       0
 4          2 Cuba     1950                            0                       0
 5          2 Cuba     1955                            0                       0
 6          2 Cuba     1955                            0                       0
 7          2 Cuba     1960                            0                       0
 8          2 Cuba     1960                            0                       0
 9          2 Cuba     1965                            0                       0
10          2 Cuba     1965                            0                       0
# ℹ 616 more rows
# ℹ abbreviated name: ¹​`Islam - Shi'a (Population)`
# ℹ 14 more variables: `Islam - Ibadhi (Population)` <dbl>,
#   `Islam - Nation of Islam (Population)` <dbl>,
#   `Islam - Alawite (Population)` <dbl>,
#   `Islam - Ahmadiyya (Population)` <dbl>,
#   `Islam - Others (Population)` <dbl>, `Islam - Total (Population)` <dbl>, …
df_islam_correct_type |>
  count(Year, Country) |> 
  arrange(desc(n)) |> 
  count(n)
Storing counts in `nn`, as `n` already present in input
ℹ Use `name = "new_name"` to pick a new name.
# A tibble: 2 × 2
      n    nn
  <int> <int>
1     1  1682
2     2   313

Remove duplicate using distinct()

df_islam |> 
  nrow()
[1] 2308
df_islam_correct_type |> 
  distinct() |> 
  nrow()
[1] 1995
df_islam_removedduplicate <- df_islam_correct_type |> 
  distinct()

Normalize Dataframe

Remove redundant ‘Islam’

df_islam_columns_renamed <- df_islam_removedduplicate |> 
  rename_with(function(x) str_remove(x, 'Islam - '),
              starts_with('Islam'))

Check Missing Value

df_islam_columns_renamed |> 
  vis_miss()

df_islam_replacedna <- df_islam_columns_renamed |> 
  mutate(across(everything(), function(x) replace_na(x, 0)))

df_islam_replacedna |> 
  vis_miss()

Check Zero Distribution (Possible Null Value)

df_islam_replacedna |>
  select(!Country) |> 
  vis_expect(function(x) x != 0)

Replace zero with NA

df_islam_replacedzero <- df_islam_replacedna |> 
  mutate(across(everything(), function(x) if_else(x == 0, NA, x)))

Check Missing Value (Second Stage)

df_islam_replacedzero |> 
  miss_var_summary()
# A tibble: 18 × 3
   variable                     n_miss pct_miss
   <chr>                         <int>    <num>
 1 Nation of Islam (Population)   1995    100  
 2 Ibadhi (Percent)               1995    100  
 3 Nation of Islam (Percent)      1995    100  
 4 Ibadhi (Population)            1985     99.5
 5 Alawite (Population)           1974     98.9
 6 Alawite (Percent)              1974     98.9
 7 Ahmadiyya (Percent)            1960     98.2
 8 Ahmadiyya (Population)         1953     97.9
 9 Shi'a (Population)             1639     82.2
10 Shi'a (Percent)                1639     82.2
11 Sunni (Percent)                1422     71.3
12 Sunni (Population)             1421     71.2
13 Others (Percent)                879     44.1
14 Others (Population)             856     42.9
15 Total (Percent)                 354     17.7
16 Total (Population)              329     16.5
17 Country                           0      0  
18 Year                              0      0  
df_islam_replacedzero |> 
  gg_miss_var()

df_islam_replacedzero
# A tibble: 1,995 × 18
   Country  Year `Sunni (Population)` `Shi'a (Population)` `Ibadhi (Population)`
   <fct>   <dbl>                <dbl>                <dbl>                 <dbl>
 1 United…  1945                   NA                   NA                    NA
 2 United…  1950                   NA                   NA                    NA
 3 United…  1955                   NA                   NA                    NA
 4 United…  1960                   NA                   NA                    NA
 5 United…  1965                   NA                   NA                    NA
 6 United…  1970                   NA                   NA                    NA
 7 United…  1975                   NA                   NA                    NA
 8 United…  1980              2407464               254121                    NA
 9 United…  1985              2186675               298183                    NA
10 United…  1990              2019386               275371                    NA
# ℹ 1,985 more rows
# ℹ 13 more variables: `Nation of Islam (Population)` <dbl>,
#   `Alawite (Population)` <dbl>, `Ahmadiyya (Population)` <dbl>,
#   `Others (Population)` <dbl>, `Total (Population)` <dbl>,
#   `Sunni (Percent)` <dbl>, `Shi'a (Percent)` <dbl>, `Ibadhi (Percent)` <dbl>,
#   `Nation of Islam (Percent)` <dbl>, `Alawite (Percent)` <dbl>,
#   `Ahmadiyya (Percent)` <dbl>, `Others (Percent)` <dbl>, …
df_islam_replacedzero |> 
  gg_miss_upset()

Check Data Constrains

Check Country Names

Exact Check
valid_countries <- codelist$country.name.en
df_country_only <- df_islam |> 
  distinct(Country)

validation_rules_country <- validator(
  Country %in% valid_countries
)

cf_country <- confront(df_country_only, validation_rules_country)
summary(cf_country)
  name items passes fails nNA error warning                    expression
1   V1   200    184    16   0 FALSE   FALSE Country %vin% valid_countries
invalid_country <- violating(df_country_only, cf_country) |> 
  arrange(Country) |> 
  pull(Country)

invalid_country
 [1] "Bosnia and Herzegovina"           "Congo"                           
 [3] "Czech Republic"                   "Democratic Republic of the Congo"
 [5] "East Timor"                       "Federated States of Micronesia"  
 [7] "German Federal Republic"          "Ivory Coast"                     
 [9] "Macedonia"                        "Myanmar"                         
[11] "Sao Tome and Principe"            "St. Kitts and Nevis"             
[13] "St. Vincent and the Grenadines"   "Swaziland"                       
[15] "Trinidad and Tobago"              "United States of America"
Automated Check by countrycode Package
tibble(invalid_country_name = invalid_country,
       converted_country_name = countryname(invalid_country))
# A tibble: 16 × 2
   invalid_country_name             converted_country_name          
   <chr>                            <chr>                           
 1 Bosnia and Herzegovina           Bosnia & Herzegovina            
 2 Congo                            Congo - Brazzaville             
 3 Czech Republic                   Czechia                         
 4 Democratic Republic of the Congo Congo - Kinshasa                
 5 East Timor                       Timor-Leste                     
 6 Federated States of Micronesia   Micronesia (Federated States of)
 7 German Federal Republic          Germany                         
 8 Ivory Coast                      Côte d’Ivoire                   
 9 Macedonia                        North Macedonia                 
10 Myanmar                          Myanmar (Burma)                 
11 Sao Tome and Principe            São Tomé & Príncipe             
12 St. Kitts and Nevis              St. Kitts & Nevis               
13 St. Vincent and the Grenadines   St. Vincent & Grenadines        
14 Swaziland                        Eswatini                        
15 Trinidad and Tobago              Trinidad & Tobago               
16 United States of America         United States
Correct Country Names
df_islam_correct_country <- df_islam_replacedzero |> 
  mutate(Country = countryname(Country))

Check Integers

validation_rules_integer <- validator(
  `Sunni (Population)` %% 1 == 0,
  `Shi'a (Population)` %% 1 == 0,
  `Ibadhi (Population)` %% 1 == 0,
  `Nation of Islam (Population)` %% 1 == 0,
  `Alawite (Population)` %% 1 == 0,
  `Ahmadiyya (Population)` %% 1 == 0,
  `Others (Population)` %% 1 == 0,
  `Total (Population)` %% 1 == 0
)
cf_integer <- confront(df_islam_correct_country, validation_rules_integer)
summary(cf_integer)
  name items passes fails  nNA error warning
1   V1  1995    574     0 1421 FALSE   FALSE
2   V2  1995    356     0 1639 FALSE   FALSE
3   V3  1995     10     0 1985 FALSE   FALSE
4   V4  1995      0     0 1995 FALSE   FALSE
5   V5  1995     21     0 1974 FALSE   FALSE
6   V6  1995     42     0 1953 FALSE   FALSE
7   V7  1995   1139     0  856 FALSE   FALSE
8   V8  1995   1666     0  329 FALSE   FALSE
                              expression
1           `Sunni (Population)`%%1 == 0
2           `Shi'a (Population)`%%1 == 0
3          `Ibadhi (Population)`%%1 == 0
4 `Nation of Islam (Population)`%%1 == 0
5         `Alawite (Population)`%%1 == 0
6       `Ahmadiyya (Population)`%%1 == 0
7          `Others (Population)`%%1 == 0
8           `Total (Population)`%%1 == 0

Check Decimals

validation_rules_decimal <- validator(
  `Sunni (Percent)` %% 1 > 0,
  `Shi'a (Percent)` %% 1 > 0,
  `Ibadhi (Percent)` %% 1 > 0,
  `Nation of Islam (Percent)` %% 1 > 0,
  `Alawite (Percent)` %% 1 > 0,
  `Ahmadiyya (Percent)` %% 1 > 0,
  `Others (Percent)` %% 1 > 0,
  `Total (Percent)` %% 1 > 0
)

cf_decimal <- confront(df_islam_correct_country,validation_rules_decimal)
summary(cf_decimal)
  name items passes fails  nNA error warning                         expression
1   V1  1995    571     2 1422 FALSE   FALSE           `Sunni (Percent)`%%1 > 0
2   V2  1995    356     0 1639 FALSE   FALSE           `Shi'a (Percent)`%%1 > 0
3   V3  1995      0     0 1995 FALSE   FALSE          `Ibadhi (Percent)`%%1 > 0
4   V4  1995      0     0 1995 FALSE   FALSE `Nation of Islam (Percent)`%%1 > 0
5   V5  1995     21     0 1974 FALSE   FALSE         `Alawite (Percent)`%%1 > 0
6   V6  1995     35     0 1960 FALSE   FALSE       `Ahmadiyya (Percent)`%%1 > 0
7   V7  1995   1116     0  879 FALSE   FALSE          `Others (Percent)`%%1 > 0
8   V8  1995   1635     6  354 FALSE   FALSE           `Total (Percent)`%%1 > 0
violating(df_islam_correct_country |> 
            select(Country, Year, contains('Percent')),
          cf_decimal)
# A tibble: 6 × 10
  Country            Year `Sunni (Percent)` `Shi'a (Percent)` `Ibadhi (Percent)`
  <chr>             <dbl>             <dbl>             <dbl>              <dbl>
1 Yemen              2000             0.69               0.31                 NA
2 Yemen People's R…  1970             1                 NA                    NA
3 Yemen People's R…  1975             1                 NA                    NA
4 Qatar              1970             0.9                0.1                  NA
5 Oman               1970             0.498             NA                    NA
6 Oman               1975             0.5               NA                    NA
# ℹ 5 more variables: `Nation of Islam (Percent)` <dbl>,
#   `Alawite (Percent)` <dbl>, `Ahmadiyya (Percent)` <dbl>,
#   `Others (Percent)` <dbl>, `Total (Percent)` <dbl>

Check Record Completeness

By Year

df_islam_replacedzero |> 
  ggplot(aes(Year)) +
  geom_line(stat = 'count') + 
  geom_point(stat='count') +
  scale_x_continuous(breaks = seq(1945, 2010, by=5))

By Country

df_islam_replacedzero |> 
  distinct(Country) |> 
  nrow()
[1] 200

Country Per Year

df_islam_year_data <- df_islam_replacedzero |> 
  count(Country) |> 
  count(n) |> 
  arrange(desc(nn)) |> 
  rename('Year Data Availibility' = n, 'Total Country' = nn)
Storing counts in `nn`, as `n` already present in input
ℹ Use `name = "new_name"` to pick a new name.
df_islam_year_data |> 
  ggplot(aes(`Year Data Availibility`, `Total Country`)) +
  geom_col() +
  scale_x_continuous(breaks = seq(1:14))

validation_rules_year <- validator(
  contains_at_least( 
      keys = data.frame(Year = seq(1945, 2010, by=5)),
      by=list(Country) )
)

cf_year <- confront(df_islam_replacedzero, validation_rules_year)
summary(cf_year)
  name items passes fails nNA error warning
1   V1  1995    868  1127   0 FALSE   FALSE
                                                                                expression
1 contains_at_least(keys = data.frame(Year = seq(1945, 2010, by = 5)), by = list(Country))
incomplete_country <- violating(df_islam_replacedzero, cf_year) |> 
  select(Country, Year)
Check Year Gap
incomplete_country |>
  group_by(Country) |> 
  mutate(lagged_Year = lag(Year, default = first(Year))) |> 
  mutate(Year_Gap = Year -lagged_Year) |> 
  ggplot(aes(Year_Gap)) +
  geom_bar() +
  scale_x_continuous(breaks = seq(0,100, by=5))

Average Starting Year
df_islam_replacedzero |> 
  select(Country, Year) |> 
  group_by(Country) |> 
  slice(1) |> 
  ggplot(aes(Year)) +
  geom_bar() +
  scale_x_continuous(breaks = seq(1945, 2010, by=5))

Average Final Year
df_islam_final_year_count <- df_islam_replacedzero |> 
  select(Country, Year) |> 
  group_by(Country) |> 
  slice_tail() |> 
  ungroup() |> 
  count(Year)

df_islam_final_year_count
# A tibble: 4 × 2
   Year     n
  <dbl> <int>
1  1970     1
2  1985     1
3  1990     4
4  2010   194
df_islam_final_year_count |> 
  ggplot(aes(Year, n)) +
  geom_col()

Normalize Dataframe

Separate Population from Percent columns

df_islam_population <- df_islam_correct_country |> 
  select(Country, Year, contains('Population')) |> 
  rename_with(function(x) str_remove(x, ' \\(Population\\)'),
              contains('Population'))

df_islam_population
# A tibble: 1,995 × 10
   Country       Year   Sunni `Shi'a` Ibadhi `Nation of Islam` Alawite Ahmadiyya
   <chr>        <dbl>   <dbl>   <dbl>  <dbl>             <dbl>   <dbl>     <dbl>
 1 United Stat…  1945      NA      NA     NA                NA      NA        NA
 2 United Stat…  1950      NA      NA     NA                NA      NA        NA
 3 United Stat…  1955      NA      NA     NA                NA      NA        NA
 4 United Stat…  1960      NA      NA     NA                NA      NA        NA
 5 United Stat…  1965      NA      NA     NA                NA      NA        NA
 6 United Stat…  1970      NA      NA     NA                NA      NA        NA
 7 United Stat…  1975      NA      NA     NA                NA      NA        NA
 8 United Stat…  1980 2407464  254121     NA                NA      NA      4975
 9 United Stat…  1985 2186675  298183     NA                NA      NA        NA
10 United Stat…  1990 2019386  275371     NA                NA      NA        NA
# ℹ 1,985 more rows
# ℹ 2 more variables: Others <dbl>, Total <dbl>
df_islam_percent <- df_islam_correct_country |> 
  select(Country, Year, contains('Percent')) |> 
  rename_with(function(x) str_remove(x, ' \\(Percent\\)'),
              contains('Percent'))

df_islam_percent
# A tibble: 1,995 × 10
   Country       Year   Sunni `Shi'a` Ibadhi `Nation of Islam` Alawite Ahmadiyya
   <chr>        <dbl>   <dbl>   <dbl>  <dbl>             <dbl>   <dbl>     <dbl>
 1 United Stat…  1945 NA      NA          NA                NA      NA        NA
 2 United Stat…  1950 NA      NA          NA                NA      NA        NA
 3 United Stat…  1955 NA      NA          NA                NA      NA        NA
 4 United Stat…  1960 NA      NA          NA                NA      NA        NA
 5 United Stat…  1965 NA      NA          NA                NA      NA        NA
 6 United Stat…  1970 NA      NA          NA                NA      NA        NA
 7 United Stat…  1975 NA      NA          NA                NA      NA        NA
 8 United Stat…  1980  0.0106  0.0011     NA                NA      NA        NA
 9 United Stat…  1985  0.0092  0.0013     NA                NA      NA        NA
10 United Stat…  1990  0.0081  0.0011     NA                NA      NA        NA
# ℹ 1,985 more rows
# ℹ 2 more variables: Others <dbl>, Total <dbl>

Check Sanity

Check whether Total value equals to all the other columns combined

Population
df_islam_population_total_check <- df_islam_population |> 
  mutate(Total_Recheck = rowSums(across(-c(Country, Year, Total)), na.rm = TRUE)) |> 
  mutate(Total_Recheck = if_else(Total_Recheck == 0, NA, Total_Recheck)) |> 
  relocate(Total, Total_Recheck, .after = Year)

validation_rules_total_population <- validator(Total == Total_Recheck)
cf_total_population <- confront(df_islam_population_total_check, validation_rules_total_population)
summary(cf_total_population)
  name items passes fails nNA error warning                          expression
1   V1  1995   1663     3 329 FALSE   FALSE abs(Total - Total_Recheck) <= 1e-08
violating(df_islam_population_total_check, cf_total_population)
# A tibble: 3 × 11
  Country      Year  Total Total_Recheck  Sunni `Shi'a` Ibadhi `Nation of Islam`
  <chr>       <dbl>  <dbl>         <dbl>  <dbl>   <dbl>  <dbl>             <dbl>
1 Burkina Fa…  2010 9.02e6       9023118 8.88e6  140256     NA                NA
2 Kenya        2010 4.20e6       4198089 3.99e6  209904     NA                NA
3 Tanzania     2010 1.55e7      15505202 1.37e7 1768358     NA                NA
# ℹ 3 more variables: Alawite <dbl>, Ahmadiyya <dbl>, Others <dbl>
Percent
df_islam_percent_total_check <- df_islam_percent |> 
  mutate(Total_Recheck = rowSums(across(-c(Country, Year, Total)), na.rm = TRUE)) |> 
  mutate(Total_Recheck = if_else(Total_Recheck == 0, NA, Total_Recheck)) |> 
  relocate(Total, Total_Recheck, .after = Year)

validation_rules_total_percent <- validator(Total == Total_Recheck)
cf_total_percent <- confront(df_islam_percent_total_check, validation_rules_total_percent)
summary(cf_total_percent)
  name items passes fails nNA error warning                          expression
1   V1  1995   1566    75 354 FALSE   FALSE abs(Total - Total_Recheck) <= 1e-08
violating(df_islam_percent_total_check, cf_total_percent)
# A tibble: 75 × 11
   Country    Year  Total Total_Recheck   Sunni `Shi'a` Ibadhi `Nation of Islam`
   <chr>     <dbl>  <dbl>         <dbl>   <dbl>   <dbl>  <dbl>             <dbl>
 1 United S…  1985 0.0104        0.0105  0.0092  0.0013     NA                NA
 2 Canada     1980 0.0072        0.0071  0.0064  0.0007     NA                NA
 3 Canada     1995 0.0043        0.0042  0.0038  0.0004     NA                NA
 4 Suriname   1975 0.196         0.196  NA      NA          NA                NA
 5 United K…  1980 0.0147        0.0148 NA      NA          NA                NA
 6 Luxembou…  2010 0.02          0.0201  0.0181  0.002      NA                NA
 7 France     2000 0.0753        0.0754  0.0669  0.0085     NA                NA
 8 Germany    2010 0.049         0.0489  0.0355  0.0034     NA                NA
 9 Yugoslav…  2010 0.031         0.0311  0.0264  0.0047     NA                NA
10 Sweden     1980 0.001         0.0009 NA      NA          NA                NA
# ℹ 65 more rows
# ℹ 3 more variables: Alawite <dbl>, Ahmadiyya <dbl>, Others <dbl>

Check Distribution

Population

 df_islam_population |> 
  plot_histogram(ncol = 1)

Percent

df_islam_percent |> 
  plot_histogram(ncol = 1)

Cleaning Summary

Verified Values

  1. All Population columns are integers
  2. All Percentage columns are decimals
  3. No missing year records in-between. For example: 1945, 1950, 1960 (missing 1955)
  4. Total Population column is equal to other Population columns combined
  5. Total Percent column is equal to other Percent columns combined

Changes Applied

  1. Removed 58 irrelevant columns
  2. Corrected 1 column’s data type
  3. Removed 303 duplicate rows
  4. Replaced all 0 values (71% of all data) with NA
  5. Corrected 16 countries’ names
  6. Separate Population and Percent columns into two dataframes.
  7. Remove ‘Islam -’ prefix and ‘(Population)’/‘(Percent)’ suffix

Other findings

  1. Extremely high prevalence of missing values.
  2. The majority of countries (60) have complete records from 1945 to 2010. But the rest vary. Ranging from only having 1 to 14 records.
  3. Despite point 1, nearly all countries have record of 2010, 2005, 2000, and so on. So, the missing records started from the bottom.

Next Steps

The plan was to analyze the growth of islam worldwide. However, the prevalenece of missing values are concerning. Most non-null values come from Sunni and Others category. So, it’s better to just use the Total column.

If you look at the data quality check at the beginning, Islam is actually the least variable with missing values. Other religions are even more dominated by missing values. This reduces the credibility of the data itself. Although, it wasn’t clearly mentioned what’s the data source anyway.

I can continue analyzing with just focusing on the overall Islam growth, but I doubt the finding would be trustworthy. For now, let’s convert the latest data frame into csv for later analysis. I’m thinking of uploading it to Tableau for easier exploration.

Export to CSV for further analysis

df_islam_population |> 
  write_csv('dataset_islam_population.csv')

df_islam_percent |> 
  write_csv('dataset_islam_percent.csv')
Back to top