Technical Assessment

This is a technical take-home test from a company for which I recently applied. The task, shown below, was to build a Rails-based, JSON-returning API from a collection of unparsed CSV documents (containing a handful of pitfalls), containing a few thousand fake health record data. Three of seven models were created for me, but database architecture / association was mine to design. The application was time-boxed to 5 hours, and additionally required me to satisfy a set of analytics requirements from the data, demanding adequate SQL/ActiveRecord query knowledge. I am very happy with my performance for the test, and am pushing up this repo as an example of my style of crafting CLEAN, MODULAR code, and preference for adherence to OBJECT ORIENTED / OOP principles. In addition to base requirements, for fun I opted to create several extra methods for additional data analysis. API endpoints were versioned, and models serialized -- also a personal choice. Sample urls found below:


  get '/hospitals', to: "api/v1/hospitals/hospital#index"
  get '/hospitals/:id', to: "api/v1/hospitals/hospital#show"
  get '/diseases', to: "api/v1/diseases/disease#index"
  get '/diseases/:id', to: "api/v1/diseases/disease#show"
  get '/medications', to: "api/v1/medications/medication#index"
  get '/medications/:id', to: "api/v1/medications/medication#show"
  get '/prescribers', to: "api/v1/prescribers/prescriber#index"
  get '/prescribers/:id', to: "api/v1/prescribers/prescriber#show"
  get '/prescriptions', to: "api/v1/prescriptions/prescription#index"
  get '/prescriptions/:id', to: "api/v1/prescriptions/prescription#show"
  get '/drugs', to: "api/v1/drugs/drug#index"
  get '/drugs/:id', to: "api/v1/drugs/drug#show"

(if you have Rails / Ruby on your local machine, just clone the repo, bundle install, rake db:migrate / rake db:seed, and rails s to launch!)

Below, find the original take-home-test spec:

===============

Setup

To setup, you will need to clone the repo down to your machine and push directly back to the remote origin.

Run the following:

bundle
rake db:create
rake db:migrate
rake db:seed

The seeds.rb file contains relevant data for the following models: Drug, Medication, and Disease.

A drug has_many medications
- For example, the drug Lipitor would have several medications of different forms and strengths.
  - Lipitor
    - Lipitor 10 mg tablet
    - Lipitor 20 mg tablet
    - Lipitor 10 mg time release capsule
    - Lipitor 20 mg time release capsule
A disease consists of only a name and an identifier.
Reference the schema or rails console for additional information about specific columns and types on these models.
You will likely need to create additional models/tables to accomplish the tasks outlined below.

Overview

We are looking for code quality, not necessarily just rushing through and checking off each task below. It is perfectly acceptable not to finish all the outlined tasks. If you are unable to finish in the time allotted we will ask you to explain your plan for implementing the rest of the tasks.

i.e. What classes, models, tables, endpoints would you have created? How would they work together?

A successful candidate will be able to demonstrate a strong grasp of data modeling, automated testing, and seperation of concerns/responsibilities.

You have been provided with a .csv file called prescription_history.csv that contains information about historical prescriptions that were written by a variety of doctors at a variety of different hospitals.

This file can be found in the highest level project directory at /prescription_history.csv

Each row contains the following columns:

prescriber_id
prescriber_name
hospital_id
hospital_name
medication_id
prescription_id
frequency_per_day
days_supply
disease

The columns medication_id and disease correspond to existing tables in the database. The id fields from the csv should be treated as industry standard identifiers that are usable at any hospital.

Tasks

Using this csv, and the corresponding data that was seeded in the database, you will need to:

Ingest the csv into the database in whatever manner seems most fitting.
- Assume this is data we purchased from a vendor in order to improve our product offering.
- Additional features will be built on top of the different elements found in the data.
Create a set of analytics around the data ingested to determine the following:
- The most commonly prescribed drug at each hospital.
- Most common medication ordered to treat each disease.
- Which prescriber appears to prescribe the greatest number of effective doses (effective_dose = one medication order * frequency_per_day * days_supply) of addictive painkillers.
  - Denoted by the opioid boolean on Drug.
  - ex: Medication A with a frequency of 2 and a days supply of 7.
    - effective_dose = 1 * 2 * 7
    - => 14
The logic related to these analytics should be tested using rspec.
Create a set of API endpoints that can be called to retrieve this information in a structured JSON format.

Submission

To submit your work, make sure that all of your work is committed and pushed. Please conclude by sending an email indicating you have finished. We already have access to the repo, as it was created for you.

We will only review any work that was committed by the agreed upon end time.

In your email, please include instructions for importing the csv and what endpoints can be used to access the relevant analytics.

thedanielvogelsang / healthcare_company_5_hour_takehome_test