Linux Capstone Project : happiness data preprocessing

You work with researchers interested in investigating changes in world happiness from year to year. Your task is to organize the data for a given year. You were provided data on country happiness from different institutes across six continents:

• Africa (AF)
• Asia (AS)
• North America (NA)
• South America (SA)
• Europe (EU)
• Oceania (OC)

To complete this task you will write a script to change the organization continent_dirs/country_files to year_dir/continent_files (cf below). This will require you to extract data for a given year from all countries and saving it in the appropriate year/continent file.

So for example if you extract all data corresponding to year 2015, your original data:

happiness_data
    ├─ AF              
    │   ├─ Algeria.csv
    │   ├─ Angola.csv
    │   └─ ...
    ├─ AS
    │   ├─ Afghanistan.csv
    │   ├─ Armenia.csv
    │   └─ ...
    ...

should become:

happiness_proj
    └─ preproc  
         └─ 2015
             ├─ AF_2015.csv
             ├─ AS_2015.csv
             ├─ ...
             └─ world_2015.csv
    

As you can see all data for a given year will be in a unique directory, with continents specified with the continent code in the filename.

The data were recorded as text files, more precisely Comma Separated Value (CSV) files, i.e. each happiness metrics for a given year are separated by a comma (,). Each line in those files represents measurements for a different year. For example the content of the happiness_data/AF/Angola.csv file is:

Country name, year, Life Ladder, Log GDP per capita, ...
Angola, 2011, 5.589000701904297, 8.945781707763672, ...
Angola, 2012, 4.360249996185303, 8.991772651672363, ...
Angola, 2013, 3.9371068477630615, 9.004610061645508, ...
...

And the cleaned / preprocessed file happiness_proj/preproc/2015/AF_2015.csv will look like:

Country name, year, Life Ladder, Log GDP per capita, ...
Benin, 2015, 3.624664306640625, 7.988193511962891, ...
Botswana, 2015, 3.761964797973633, 9.608841896057129, ...
Burkina Faso, 2015, 4.4189300537109375, 7.562853813171387, ...
Cameroon, 2015, 5.037964820861816, 8.180439949035645, ...

To accomplish your data cleaning / preprocessing your script will rely on using the program get_year located inside the bin directory within your home dir (e.g. at /home/ada/bin/get_year) which extracts the row corresponding to a specific year.

Furthemore you were told to exclude datasets from countries with less than 5 time points (i.e. measurements for fewer than 5 different years) are those were deemed not suitable to study longitudinal changes.

PART 1

---

Reorganizing the data and defining the script main variables.

1. Create a directory happiness_proj inside the ada directory

2. Copy recursively the content of happiness_data to a directory raw inside happiness_proj (note: happiness_data is in the ada directory). The results should be:

happiness_proj
    └─ raw
        ├─ AF              
        │   ├─ Algeria.csv
        │   ├─ Angola.csv
        │   └─ ...
        ├─ AS
        │   ├─ Afghanistan.csv
        │   ├─ Armenia.csv
        │   └─ ...
    ...

3. Create a new script called preproc_year_v1 inside the happiness_proj directory, that:

• assigns to a variable IN_DIR the absolute path to your happiness_proj directory

• assigns to a variable RAW_DIR the path to the raw directory, using the value of the IN_DIR variable previously defined

• assigns to a variable YEAR an integer between 2008 and 2021 (choose it yourself)

• assigns to an array CSV_FILES (do not forget quotes " ") all the CSV files which:

  • are in the subdirectories of ${RAW_DIR}
  • ends with the .csv file extension

  TIP: find the right “globbing” expression and do not forget parentheses to define the array

• Within a SINGLE loop, loop on each item of the resulting array ${CSV_FILES[@]} and print:

  • the number of time points in that file
    • use cat, | and wc -l within a command expansion $()to save the number of time points in a variable N_TIME_POINTS, then display the variable value
  • the name of the file (e.g. Armenia.csv) to save in a variable COUNTRY_FILENAME [TIP: use one string manipulation statement, based on the # symbol]
  • the name of the file parent directory (e.g. AF) to save in a variable CONTINENT_CODE [TIP: use two string manipulation statements: one based on the % symbol to create a temporary variable, and another one based on the # symbol applied to that temporary variable]

  NOTE: you only need one loop to do everything, so do not create more than one loop

4. Make the script executable and run it. Use shellcheck preproc_year_v1 to check for errors

PART 2

---

Preparing the use of the get_year command, without running it yet in the script.

1. Look at the usage of the command get_year located in your home bin directory (e.g. at /home/ada/bin/get_year) by typing the command without arguments
2. If which get_year doesn't return the path to that program, this means the directory containing get_year is not in your shell global PATH variable. In this case, to avoid having to type the full path to the get_year program every time you need it:

• make a backup copy of your account .bashrc file (e.g. cp ~/.bashrc ~/.bashrc_backup)
• edit your .bashrc file so that to add the path to the directory containing get_year to your PATH variable
• update your shell environment by sourcing your .bashrc file
• test that get_year is now reachable by your shell (i.e. that it is in your PATH variable) by typing which get_year
  • If you are working on a Mac and which get_year does not work even after updating your .bash_profile, try adding the command source ${HOME}/.bashrc to the .profile file in your home directory, save it, and then start a new terminal to test if it works. Report to Slack if you are stuck here.

3. After reading the usage of get_year, test it on a test output file to check it works. Then delete the test output file.
4. Save preproc_year_v1 as preproc_year_v2 and edit is so that it also:

• assigns to a variable OUT_DIR the path to a directory named with the year chosen and inside a preproc directory itself within happiness_proj (e.g. /home/ada/happiness_proj/preproc/2018 if the year is 2018) [TIP: define OUT_DIR using the IN_DIR and YEAR variables]

• create the OUT_DIR directory if the directory does not exist

  TIPS:

  • use an if-else statement with a unary condition
  • use the -p option of mkdir if you need to create new directories when one is inside the other

• loop on each item of ${CSV_FILES[@]} and:

  • if the file has less than 5 time points, print a message indicating the file is excluded

    TIP: use the variable to which you assigned the number of time points in an if-else statement

  • if the file has 5 or more time points:

    • print a message indicating the file is included for preprocessing
    • define a variable OUT_FILE with what should be the path where to save the data extracted for that day, e.g. the data from AS/Armenia.csv extracted for year 2018 should be saved in the file happiness_proj/preproc/2018/AS_2018.csv [TIP: use variables OUT_DIR, CONTINENT_CODE and YEAR previously defined to define OUT_FILE]
    • print to the screen the value of OUT_FILE

  • print the command get_year which should be used to extract the data at the chosen year with the appropriate arguments (i.e. echo "get_year -f ...") [TIP: make sure you have the right CSV file, output file and year, and that you do not forget to use double quotes " " appropriately (some CSV files have space in them !)]

5. Make the script executable and run it. Use shellcheck preproc_day_v2 to check for errors. Check the printed commands make sense.

PART 3

---

The script is almost ready to be run with get_year. However when you tested that program (part 2 task 3), you may have noticed that there is no file header (i.e. the name of the columns are not in the output file). Let's address this.

1. Save preproc_day_v2 as preproc_day_v3
2. All the CSV files have the same header so we can just take the first one to extract that header. After the CSV files array definition but before the for loop:

• create a variable FIRST_CSV having for value the path to the first CSV file (TIP: use index 0 to extract the first item of the array)
• create a variable HEADER having for value the first line of that first CSV file [TIP: use command expansion $() to save the output of a command mixing cat, | and head to extract the first line]

3. Edit the script so that the printed get_year command will be run instead of printed (i.e. just remove the appropriate echo)
4. Modify your script so that before running get_year two different things will happen depending if the output file already exists or not:

• if the output file doesn't exit, write the header to it [TIP: use echo with the output redirection >] before running get_year
• if the output file exists, then simply run get_year (the header already exists so no need to add it)

5. Make the script executable and run it. Use shellcheck preproc_day_v3 to check for errors. Check the output is as described in the project introduction.
6. Do not hesitate to delete your output directory if needed to re-test the script, but please be extremely careful when using the rm command on the terminal (you will not be able to recover your files if you make a mistake). Instead of rm -R <directory>, you can use rm -I -R <directory> to be prompted before confirming file deletion (cf manual with man rm).

PART 4

---

Create a file including the data for all continents.

1. Save preproc_day_v3 as preproc_day_v4
2. Create a variable WORLD_OUT_FILE having for value the name of the file to receive the data from all continents. It should be of the form happiness_proj/preproc/2018/world_2018.csv [TIP: use variables OUT_DIR and YEAR previously defined]. Since this variable is independent of the continent name it should be defined before the for loop.
3. Adding data to the world output file relies on executing a second time get_year, and adding the header if WORLD_OUT_FILE doesn't exist yet (TIP: repeat the check to add the header and repeat the call to get_year in the appropriate place without forgetting to use WORLD_OUT_FILE instead of OUT_FILE)

PART 5

---

Allow the script to parse user inputs with getopts.

1. Save preproc_year_v4 as preproc_year_v5 and use getopts for the user to indicate the raw data directory RAW_DIR, the YEAR and the output directory OUT_DIR on the command line so that:

• RAW_DIR and OUT_DIR are compulsory arguments (the program should exit with error code 1 if the -i or -o option respectively are not indicated on the command line by the user)
• YEAR is optional and has default value 2010 if not indicated on the command line

  • YEAR should be between 2010 and 2021 (program should exit if this is not the case)

    TIP: to test two conditions, you can use AND or OR logical operators, which are respectively && and ||, for example to test that a variable name is not equal to C3PO or R2D2 one could write: if [[ ${name} != "R2D2" && ${name} != "C3PO" ]]; then ... (please do not forget the space on each side of all the operators, as well as after [[ and before ]])

2. Edit the program so that it displays a usage function (indicating optional argument in square brackets) when:

• the script is called without arguments

• the script is called with improper parameters (e.g. YEAR=1900)

  TIP: copy/paste the content of /home/ada/getopts_skeleton or /home/ada/getopts_skeleton_full and then adapt it

• Test the script