Candidate Instructions
User Manual:
Open the PDF directly: View PDF 
.
Page Count: 2
Candidate instructions: Hired/Not-Hired
classifier for Pokemon Trainers
You are asked to give advice for teams of Pokemon Trainers all across Pokemon World who are
looking for new trainers to add to their team.
Trainers fall into one of 8 classes (Curmudgeon
, Doctor
, Dragon
Tamer
, Engineer
, Nurse
,
Pokemon
Ranger
, Scientist
and Skier
). In each region of Pokemon world, and for each Trainer
Class, there’s a distinct team that’s hiring (regions are Kanto
, Johto
, Hoenn
, Sinnoh
, Unova
,
Kalos
, Alola
and the Sevii
Islands
). A team can be uniquely identified by trainer class and
region.
Given data from job applications for Pokemon Trainers:
1. Which team (i.e. combination of region and trainer class) is the most competitive (meaning
the most difficult to get hired into)?
2. Train a model that predicts whether candidates will get hired or not.
3. If you were a team hiring a Pokemon Trainer, how would you use this data to make the hiring
process easier?
Please aim to spend no more than 3 hours on the entire exercise and explain your thought
process and any choices you make, such as how to deal with missing values and which
columns to use, as well as how you evaluate your model.
Please submit all the files necessary to run you code (for example a Jupyter notebook with
Python code or an R markdown file), together with any instructions on how to run it, including
which packages must be installed.
Data
The applications of aspiring Pokemon Trainers can be found in the file
pokemon_trainer_application_data.csv
Columns
- Target column: hired (0 or 1, where 1 means "hired")
- Columns indicating team are a combination of region in Pokemon world and Trainer class:
(PokemonWorldRegion, PokemonTrainerClass)
Furthermore, there are many more columns, of which you should decide whether you want to
use them or not. Some have to do with a trainer's previous education or training experience, or
which Pokemon they have won badges with (for instance GymBadge1Pokemon). Other columns
may contain details of the position for which they're applying (for instance
PositionForTrainingPokemon which states which Pokemon they will be training), or details
about the application itself (for instance ApplicationType or ApplyDate).
You don’t need to know anything about Pokemon to do this exercise, but for your information,
there are three types of Pokemon-specific names. These are present in the following columns:
Region names appear in columns: PokemonWorldRegion
Trainer classes appear in columns: PokemonTrainerClass
Pokemon names appear in columns: CurrentPokemonTraining,
PositionForTrainingPokemon, GymBadge1Pokemon, …, GymBadge4Pokemon. A list of
Pokemon names can be found in the file pokemon_names.csv
