Data Science for Beginners: Data Science Intro Course

Frame a case study for a U.S. health insurer to allocate a limited budget and maximize quality care by clarifying goals and readmission risks after acute myocardial infarction.

Apply data understanding by computing descriptive statistics, examining variable distributions and correlations, and verifying data quality with experts, possibly revisiting data collection to add secondary diagnoses for acute myocardial infarction.

Explore data preparation to turn understanding into action by cleaning data, handling missing values, normalizing and standardizing features, and creating new features to improve readmission prediction.

Split the data into a training set to build models predicting readmission. Compare classification algorithms like logistic regression and decision trees, emphasizing interpretability for actionable insights.

Learn how to evaluate classification models using the confusion matrix and metrics like accuracy, precision, and recall; understand false negatives, bias-variance tradeoffs, and the train/test split.

Deploy readmission risk models for end users like insurance companies, using dashboards and BI tools updated daily to drive follow-up, medication adherence, and targeted resource allocation.

Discover machine learning in action with the IBM Watson assistant, a machine learning as a service chatbot, and learn how data scientists collaborate with developers to automate basic data tasks.

Explore how chatbots, virtual assistants, and virtual agents transform customer service with 24/7 availability. From pizza orders to banking, data science enables natural, helpful responses and diagnosis and treatment advice.

Sign up for a free IBM Watson Assistant account on IBM Cloud to get started for free and build your chat bot; existing users can log in.

Learn to build a health care chatbot with a Watson assistant by creating an assistant, adding a dialogue skill, and configuring welcome and anything else nodes to handle user input.

Create an intent like book_appointment, add user examples, and train Watson Assistant Lite to recognize bookings. Connect the intent to a dialogue and test training for correct understanding.

Create an appointment type entity with values like medical care, dental, and psychological, add synonyms and fuzzy matching, and connect it to dialogue to identify requests and train the model.

Explore the learning center and documentation to learn about building complex dialogues with slots, digressions, and guided tours, then apply tutorials on creating entities.

Explore natural language processing as a service to turn unstructured text into structured data for machine learning, with text classification and intent detection.

Explore Python libraries for numerical computing, data cleaning, and visualization—NumPy, Pandas, Matplotlib, Seabourne. Learn machine learning with Saikat learn and Karash, plus tensor flow; explore R libraries and SAS base.

Explore how restful APIs use HTTP to send and retrieve data between applications. Data scientists access remote data via Python requests, receiving XML or JSON for analysis and model scores.

Define data sets as structured data in tabular formats like Excel or CSV, with rows as observations and columns as variables, and note open data sources, licenses, and creation methods.

Explore github as a code hosting platform for version control and collaboration, track changes, review code, add comments, suggest changes, assign tasks, and back up your work.

Create a repository to organize a single project with folders, files, and a readme. Set privacy, upload content, and learn the basics as branching is covered in the next lecture.

Discover how to create branches and commit changes, using the main production branch and feature branches within the GitHub flow.

Learn to create and review pull requests on GitHub, compare branches, view diffs, and merge changes from a readme edits branch into the main branch, then delete the branch.