Richard Davies

An overview of my Data Science course, developed with Denes Csala and colleagues in the Economics Observatory Data Team.

If you are not a current student, but like the sound of the course, get in touch with the team at office@economicsobservatory.com. We will be running a course for mid-career professionals in 2025.

Course overview

1. Hosting Visualisations: A new approach.
   • Motivation.
   • Resources: new tools.
   • Building blocks: HTML, CSS, JavaScript.
   • Charts as data.
2. Debating. Making your point with data.
   • Good, Bad and Ugly – when charts go wrong.
   • DataViz heroes.
   • Principles to help you avoid ChartJunk.
   • Black Hat, White Hat – two types of adjustment.
3. Structure. The Language of Data.
   • Visualization as a sign system.
   • Visual encoding.
   • Data guidelines.
   • Data sources.
4. Data.
   • Data collection, data storage, data access.
   • Manipulating Data.
   • APIs.
   • Scrapers.
5. Programming.
   • Flow Control in Data Science.
   • Conditionals.
   • Loops.
   • Dashboards.
6. Break – Project Work Week.
7. Maps.
   • Cartography history: first choropleth to modern maps.
   • Projections.
   • Base maps.
   • Putting data on maps.
8. Big Data
   • Intro to Big Data: history and revolution.
   • Definitions, development and usage.
   • Administrative data: impact and safe usage.
   • Ethics of big data.
9. Analytics.
   • Comparison. Rebasing, de-trending, normalizing.
   • Distribution, correlation and regression charts.
   • Events and DiD charts.
   • Forecasts, PCA and sentiment analysis.
10. Machine Learning.
    • ML definitions and history.  
    • Unsupervised. Clustering and dimensionality reduction.
    • Supervised. Regression and classification.
    • Model evaluation and hyperparameter tuning.  
11.  Large Language Models (LLMs)
    • Data gathering.
    • Data cleaning.
    • Data visualisation.
    • Audit

Lecture 1. Motivation, and getting set up

The aim of the first week is to get you set up as a coder. The course assumes zero prior coding knowledge. Within a couple of hours you will be confidently running HTML, CSS and JavaScript on your own live web page.

For some preliminary motivation and ideas, the links below give a compendium of types of charts, and some links to previous students’ web pages.

• https://rdeconomist.github.io/library
• https://RDeconomist.github.io/datascience2022

To get started, in our first lecture and practical we meet three types of code:

• HTML. https://developer.mozilla.org/en-US/docs/Web/HTML
• CSS. https://developer.mozilla.org/en-US/docs/Web/CSS
• Javascript. https://developer.mozilla.org/en-US/docs/Web/JavaScript

And we will start using four tools.

• Vega-Lite. https://vega.github.io/vega-lite/
• GitHub. https://github.com/
• Visual Studio Code. https://code.visualstudio.com/
• ECO Data Hub. https://www.economicsobservatory.com/create

By the end of lecture 1, you will know what all of these things are, and will be using them to display data in your own website.