Module 8: Modern Analytics

All modules 3 sessions · 13 hours

  1. Reproducible Reporting in Python with Quarto and xlsxwriter

    6 hrs

    Session lead: Sammi Rosser

    In this session, we spend a bit of time learning about Quarto, allowing you to create good-looking reports that weave together code, text, images and more in a neat, easily-distributable format. We also have a look at using a Python script to interact with the command line and produce multiple outputs from a single Quarto file using different parameters.

    We then briefly dive into the world of xlsxwriter - a handy tool for automatically producing Excel spreadsheets from your Python code.

    This session makes use of the hsma_reproducible_reporting environment

    Slides Video Repository

  2. Time-Series Forecasting

    6 hrs

    Session lead: Elliot Coyne, Tom Monks

    Time-series forecasting methods are a useful thing to have in your analyst toolbox. In this session, we cover how to prepare data for forecasting, how to train simple models, how to assess your forecasts, and how to use the modern forecasting library Prophet. We also discuss when time series forecasting methods are a good fit for your problems, and when you should consider turning to other approaches.

    This session has not yet been fully rewritten for HSMA 6.

    However, you can watch the sessions from HSMA 5, which cover the same material.

    Some bonus videos from previous rounds of HSMA are also linked below.

    Due to some troublesome environment issues, the exercises are not currently completely functional. We would instead recommend Tom Monk's excellent exercises, on which the HMSA 5 exercises were originally based. Tom's exercises have been updated to work in 2024 and are used on the Exeter Health Data Science MSc.

    Slides - Introduction to Forecasting Principles and Naive Models Video - Introduction to Forecasting Principles and Naive Models Slides - Forecasting with the Prophet Library Video - Forecasting with the Prophet Library Recommended Exercises Bonus Video - Auto-regressive Forecasting using OLS Bonus Video - Introduction to using Feedforward Neural Networks for Forecasting Bonus Video - Introduction to forecasting using Feedforward Neural Networks using Keras and TF

  3. Tips, Tricks, and the Things We Couldn't Cover

    1 hr

    Session lead: Sammi Rosser

    In this slide-only session, we go over some useful things you may want to consider in your own coding projects, as well as briefly touching on a range of topics that we couldn't cover in the course. It's useful to know of these approaches so you are aware of the possibilities and can spot situations in which these may be a more appropriate approach than the ones we've had time to cover in detail on the course. Using your Python knowledge and general data science and operational research concepts, you are now well-placed to dig further into any topics in this session that catch your eye.

    Slides
No matching items