Exercises/Ex.1 - Data Science with Jupyter.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src = http://i1377.photobucket.com/albums/ah57/drmingle1/Jupyter%20-%20Healthcare%20Data%20Science/BosDataFestival2015_zpsqpbcvzw8.png alt=\"Matrix\" style=\"width: 600px;\"/>\n",
    "\n",
    "\n",
    "# Finding Type II Diabetes In A Patient Population\n",
    "Given a year of EHR data for patients without Diabetes, we predict which patients will be diagnosed with Diabetes in the next year."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***************************\n",
    "#Excercise 1\n",
    "\n",
    "Choose between A and B.\n",
    "***************************"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Data Scientist A**\n",
    "\n",
    "Company wants to predict patients with diabetes using EMR data.\n",
    "Company feels this will reduce cost significantly.\n",
    "\n",
    "We are going to use a logloss metric. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "**===================================================================================================**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Data Scientist B**\n",
    "\n",
    "#PHASE I: Business Understanding\n",
    "\n",
    "####Determine Business Objectives\n",
    "1. *Background*\n",
    "    * **Organization**:\n",
    "        * ACME Healthcare\n",
    "    * **Problem Area**:\n",
    "        * Clinical Diagnosis\n",
    "        * As of 2014 there are 8.1 million people or 27.8% of people with diabetes who are undiagnosed. Total costs (both direct and indirect) is estimated to be $245 billion.\n",
    "        * Data Science needs to be promoted as a key way to assist the business.\n",
    "        * The organization is motivated to determine the patient population which has a high-likelihood of type II diabetes in order to offer better care for patients as well as reduce cost of care.\n",
    "        * Target groups we will work with are: chief medical officer, biostatisticians\n",
    "        * The organization is expecting to make use of a high-accuracy model to engage with physicians about their population\n",
    "    * **Current Solution**: \n",
    "        * *Description*: Hospital has rank ordered their patients based on a series of rules that have been identified by clinicians. The hospital is taking the top 5 percent of the patients who fall into that category and those cases are being reviewed by a physician to determine likelihood of being an underdiagnosed type II diabetic. \n",
    "        * *Advantages/Disadvantages*: Since the solution is more conventional the hospital feels more secure in their findings, although the process is prone to human error.  However, the current process is costly since it involves physicians and there is a significant lag in results as it takes six months to complete a single-iteration.\n",
    "        \n",
    "2. *Business objectives*\n",
    "    * Primary Objective: Given a year of EHR data for patients without Diabetes, aim to predict which patients will be diagnosed with diabetes in the next year.\n",
    "    * Other Objectives: What features are most predictive? The business does not want to create unnecessary alarm for their patient population. \n",
    "    * Benefits: The hospital expects that they will considerably reduce cost for the hospital by systematically discovering patients who may have fallen through cracks with regard to type II diabetes.\n",
    "    \n",
    "3. *Business success criteria*\n",
    "    * A successful and useful outcome for this Data Science project from the hospitals point of view is that they will experience their HCAP scores increase. Improved rankings bode well for the hospital. The potential for increase quality of care will aid the hospital as the healthcare environment swithces from fee-for-service to value-based care. \n",
    "    * Clinicians will assess the success of the program\n",
    "    * The hospital CFO will review the daily operations, budgeting, analytics, and oversight of expenses as it relates to this project. They will be looking for a 7 percent drop in cost of care within the first 12-months of deploying this model.\n",
    "    * The payers Senior Director, Sales & Marketing will determine the profitable growth and development of increasing patient population through marketing programs around this project. \n",
    "    * The hospital VP, Quality Improvement will determine the value of implementing this Data Science project as a clinical program designed to improve patient health outcome through promotion of preventive care services and management of chronic conditions. \n",
    "     * VP, IT Quality Assurance & Compliance will determine how well this model can be monitored and maintained once in production. \n",
    "    * The business has the aim of predicting the probability that each person has a diagnosis of Type 2 Dieabetes Mellitus which will be evaluated using the log loss metric.\n",
    "\n",
    "$$ log loss = \\frac{1}{n}\\sum_{i=1}^{N}y_{i}log(\\hat{y}_{i})+(1-y_{i})log(1-\\hat{y}_{i}) $$\n",
    "\n",
    "Where:\n",
    "$$N$$  is the number of patients \n",
    "$$log$$ is the natural logarithm\n",
    "$$\\hat{y}_{i}$$ is the posterior probability that the ith patient has diabetes\n",
    "$$y_{i}$$ is the ground truth ($$y_{i}=1$$ means the patient has diabetes, $$y_{i}=0$$ means that the patient does not).\n",
    "\n",
    "######Assess Situation\n",
    "1. Inventory of resources\n",
    "    * Personnel Sources:\n",
    "        * Identified project sponsor: Chief Medical Officer\n",
    "        * Identified technology personnel (all available over the next 60-days as needed):\n",
    "            * Chief Technology Officer\n",
    "            * Senior Data Architect\n",
    "        * Identified support personnel (all available over the next 60-days as needed):\n",
    "            * Business analyst\n",
    "            * Data Scientist\n",
    "            * Statisticians\n",
    "            * Medical Coding Expert\n",
    "            * Clinician\n",
    "            * Project Manager\n",
    "    * Sources of Data and Knowledge: \n",
    "        * Identified data sources:\n",
    "            * Electronic Health Records (EHR) which includes information on diagnoses, lab results, medications, allergies, immunizations, vital signs, and health behaviour. \n",
    "        * Identified types of data sources:\n",
    "            * 15 types of EHR as an extract\n",
    "        * Identified knowledge sources:\n",
    "            * Clinicians\n",
    "            * Analysts\n",
    "            * Data Scientist\n",
    "            * Data Architect\n",
    "            * Current model\n",
    "        * Identified types of knowledge sources:\n",
    "            * Online sources\n",
    "            * Written documentation\n",
    "            * Experts\n",
    "        * Describe the relevant background knowledge: \n",
    "            * One of the two major types of diabetes, the type in which the beta cells of the pancreas produce insulin but the body is unable to use it effectively because the cells of the body are resistant to the action of insulin. Although this type of diabetes may not carry the same risk of death from ketoacidosis, it otherwise involves many of the same risks of complications as does type 1 diabetes (in which there is a lack of insulin).\n",
    "\n",
    "            * The aim of treatment is to normalize the blood glucose in an attempt to prevent or minimize complications. People with type 2 diabetes may experience marked hyperglycaemia, but many do not require insulin injections and can be treated with diet, exercise, and oral hypoglycemic agents (drugs taken by mouth to lower the blood sugar).\n",
    "\n",
    "            * Type 2 diabetes requires good dietary control including the restriction of calories, lowered consumption of simple carbohydrates and fat with increased consumption of complex carbohydrates and fiber. Regular aerobic exercise is also an important method for treating both type 2 diabetes since it decreases insulin resistance and helps burn excessive glucose. Regular exercise also may help lower blood lipids and reduce some effects of stress, both important factors in treating diabetes and preventing complications.\n",
    "\n",
    "            * Type 2 diabetes is also known as insulin-resistant diabetes, non-insulin dependent diabetes, and adult-onset diabetes.\n",
    "        \n",
    "    * Computing Resources:\n",
    "        * Hardware platforms\n",
    "            * Amazon EC2 Instance:C4 instance which offer a compute-optimized instance, featuring the highest performing processors and the lowest price/compute performance in EC2. \n",
    "\n",
    "Model|vCPU|Mem (GB) |Storage| Dedicated EBS Throughput (Mbps) \n",
    "-|-|-|-\n",
    "    c4.4xlarge|16|30|EBS-Only|2,000\n",
    "    \n",
    "        * Software: \n",
    "            * R\n",
    "            * Python\n",
    "            * Jupyter\n",
    "        \n",
    "2. Requirements, assumptions & constraints\n",
    "    * Verify Constraints\n",
    "\n",
    "Constraint |Result\n",
    "-|-\n",
    "    All usernames/passwords tested for data access?| YES\n",
    "    Verified all legal constraints on data usage? | YES\n",
    "    Financial constraints covered in the project budget?| YES\n",
    "\n",
    "\n",
    "Task|Duration|Running Total\n",
    "-|-|-|-\n",
    "    Describe Data | 10 Days | 10 Days\n",
    "    Explore Data | 15 Days | 25 Days\n",
    "    Inferential Statistics | 10 Days | 35 Days\n",
    "    Modelling | 20 Days | 55 Days\n",
    "    Write-Up | 5 Days| 60 Days\n",
    "    \n",
    "    * Comprehensibility and quality of results\n",
    "    * Security\n",
    "    * Legal issues\n",
    "    * Assumptions\n",
    "    * Constraints\n",
    "3. Risks and Contingencies\n",
    "    * Risks: not being able to get the data, experts, or technology needed\n",
    "    * Contingency plan: \n",
    "        * If live data is not available, we will move forward with data extract only.\n",
    "        * If clinicians are not available to help with domain expertise because of competing priorities the data science team will pull together cursory research on relevant topics to gain intuition about the domain.\n",
    "        * If EC2 instance usage is exceeded we will migrate to workstations to continue solution development. \n",
    "4. Terminology\n",
    "    * Relevant Business Terminology:\n",
    "        * **ICD9** - The International Classification of Diseases, Ninth Revision.\n",
    "        * **Diabetes Mellitus** - is a chronic, lifelong condition that affects your body's ability to use the energy found in food. \n",
    "        * **Type II** - Begins with insulin resistance, a condition in which cells fail to respond to insulin properly. As the disease progresses a lack of insulin may also develop. \n",
    "    * Relevant Data Science Terminology:\n",
    "        * **Log Loss** - this is the loss function used in (multinomial) logistic regression and extensions of it such as neural networks, defined as the negative log-likelihood of the true label given a probabilistic classifier's predictions. \n",
    "        * **GBM** - (Gradient Boosted Machines) is a machine learning technique for regression and classification problems, which produces a prediction model in the form of an ensemble of weak prediction models, typically decision trees.\n",
    "        * **RF** - (Random Forest) the combination of bagging, subspace sampling, and decision trees.\n",
    "5. Costs and Benefits\n",
    "    * Cost of the project: This project will cost 600,000 for the first 90-days and then 15,000 month for maintenance of the model.\n",
    "    * Potential benefits if successful: The hospital will been seen as a leader innovation by using data science to extend what the physicians are already doing. \n",
    "\n",
    "######Determine Data Science Goals\n",
    "1. Data science goals\n",
    "    * To build a model that takes in de-identified data of patients electronic health records and determine which patients have a diabetes diagnosis, as defined by ICD9 codes 250.**.\n",
    "2. Data science success criteria\n",
    "    * Predictive Accuracy: logloss less than 0.40\n",
    "    \n",
    "######Produce Project Plan\n",
    "Describe the intended plan for achieving the data science goals and thereby the business goals. Specify the steps to be performed during the rest of the project, including the initial selection of tools and techniques.\n",
    "1. Project plan\n",
    "\n",
    "Stage|Duration |Resource Required | Inputs | Outputs| Dependencies\n",
    "-|-|-|-|-\n",
    "    Describe Data | 10 Days | Data Scientist | SQL Tables from EMR | Summary Statistics; Visualizations | Data Architect\n",
    "    Exploratory Data Analysis |15 Days | Data Scientist, Clinician| Data | Outliers; Visualizations | Data Description\n",
    "    Inferential Statistics | 10 Days | Data Scientist, Clinician|Data |Imputation of Values|Clean Data Set\n",
    "    Feature Engineering |1 Day| Data Scientist| Data| Enhanced Features| Inferential Statistics\n",
    "    Feature Selection| 1 Day| Data Scientist| Data | Important Features | Feature Engineering\n",
    "    Model Training | 15 Days|Data| Data Scientist| Highest Accuracy Models | Feature Selection\n",
    "    Evaluation| 4 Days | Models| Data Scientist | Identify Successful Models | Model Training\n",
    "    Ensemble | 4 Days | Data Scientist |Evaluation; Models| Stable Model| Evaluation\n",
    "    \n",
    "    \n",
    "2. Initial assessment of tools and techniques\n",
    "    * Data Science Tool Selected:\n",
    "        * R for descriptive statistics and exploratory analysis\n",
    "        * R for model development\n",
    "        * Jupyter for presentation of findings"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.4.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}