A package for R language that will bestow users with magical data analytic capabilities. This is mainly to allow researchers at the Cambridge Prosociality and Well-being Lab to focus more on collaborative research and less on doing silly gruntwork.

The main point of magic is to remove the complexity of connecting to a database by hiding it from view: you set things once and the rest is magic.

The easiest way to get some magic on your computer, please type the following to your R console:

    install.packages("devtools")
    require("devtools")
    
    install_github("magic", "vanderlowe")
    require("magic")

Please go to http:https://dev.mysql.com/downloads/connector/odbc/#downloads and download the driver for Windows platform (depending on the version of Windows you have). For any computer that is less than two years old, you probably should install the x64-bit version.

1. Click start and type ODBC.
2. Something called Data Source (ODBC) should appear. Click on this.
3. Now click the System DSN tab.
4. Select Add... in this tab.
5. Choose MySQL ODBC 5.2w Driver
6. Fill in the following values on the new screen:
7. In the "Data Source Name:" field, type: CPW Server (Leave the description field blank.)
8. For the "TCP/IP Server field:", enter: koganserver001.psychol.cam.ac.uk (Leave the "Port:" as 3306)
9. For the "User:" Field, enter your Hermes user name. Your Hermes user name is the beginning of your Cambridge email address, such as ak823.
10. For the value of "Password", enter the password you received from Ilmo. (Leave the "Database:" field blank.)
11. Click OK!

Most of your R scripts that use the lab data should start with:

    require(magic)

Including this in your scripts will make the useful functions in magic available to you. These include: magicSQL, magicExplain, and magicFindVariable.

The function you will use the most is magicSQL. You can try it by typing:

    magicSQL()  # Don't forget the parentheses!

Without any further instructions, magicSQL shows you the databases that are available to you.

Let's see the tables in the cpw_iOpener database. You can investigate the database by typing:

    magicSQL(db = "cpw_iOpener")

This shows that the cpw_iOpener database has only one table, iPPQ. To see what variables are in the iPPQ table, type:

    magicSQL("SHOW FIELDS FROM iPPQ", db = "cpw_iOpener")

To create a dataset that includes gender, age_group, and love_job variables, you type:

    mydata <- magicSQL("SELECT gender, age_group, love_job FROM iPPQ", db = "cpw_iOpener")

Voilà.

Many datasets are complex and finding variables by listing everything is tedious. The function magicFindVariable helps you to quickly narrow down the variables of interest.

Let's say we are interested in close relationships, especially marriage. To find variables that might have to do something marriage, you can type:

    magicFindVariable("marriage")

This function looks at the variable descriptions and tries to find mentions of the word "marriage" in them.

Based on the results, the variable f188 seems especially interesting, so let's find out more about it.

    magicExplain("f188")

If we were to use this variable in our analyses, the function even gives the command for downloading the data.

Below is a more complex example script to work with the myPersonality data.

    # Every journey starts with a single step:
    require(magic)
    
    # Which tables does the cpw_myPersonality database have?
    magicSQL(db = "cpw_myPersonality")
    
    # Which variables do the tables big5 and address have?
    magicSQL("show fields from big5", "cpw_myPersonality")
    magicSQL("show fields from address", "cpw_myPersonality")
    
    # How many observations do variables ope and location_lat have?
    magicSQL("SELECT COUNT(ope) FROM big5", "cpw_myPersonality")
    magicSQL("SELECT COUNT(location_lat) FROM address", "cpw_myPersonality")
    
    # How long does it take to download data that joins the two?
    system.time(nature <- magicSQL("SELECT big5.ope, address.location_lat, address.location_lon, address.hometown_lat, 
    address.hometown_lon FROM big5, address WHERE big5.userid=address.userid", "cpw_myPersonality"))
    
    # What does plotting the location data reveal?
    plot(nature$location_lon, nature$location_lat)

There are a number of helpful magic functions at your disposal, especially for working with country-level data.

    # Get ISO 3166-1 country codes
    iso2("Zambia")
    iso3("Zambia")
    
    # Get a list of country data
    countries()

If you use magic in your work, make sure to cite it:

    citation("magic")