Introduction. Course schedule and expectations. Why do data visualization? Historical examples. Exploratory vs. explanatory. A visualization analysis framework.
Big picture questions. Why are computers and humans involved in data visualization? Why show the data? Why allow people to interact with it? What resource limitations are there? Also, course admin.
Dataset types — tables, charts, networks, fields, geometry, hiereachies. Data scales — nominal, ordinal, interval, ratio. Data ordering — sequential, diverging, cyclic.
A refresher on web core technologies – HTML / CSS / JS. Useful JS design patterns that you will see over and over again. Function chaining, interfacing with backend systems with JSON and AJAX, dealing with asynchronous issues with callbacks and promises.
Pull out data.gov.sg's PSI API data and populate it into a simple HTML table. Prep for future assignments / the project.
Why are people using data viz? We'll analyze by looking at a high-level framework of actions and targets. Actions: Analyze, search, query, etc. Targets: Trends , outliers , features, etc.
Anand is the CEO of Gramener, a data analytics company that just got sold to Straive (SPi Global) last year. Straive is estimated to have a revenue of $1.9 billion in 2023.
Anand will talk about visualizing large language model hallucinations, i.e. the process of creating a visual explorer for LLM token exploration.
He is addicted to MInecraft (thanks to his daughter) and you can find a list of his talks here.
4 levels of visualization validation: domain/problem → data/task abstraction → viz idiom → algorithm / technology. Case studies and examples.
Alice Lee is an assistant director at Singapore Food Agency(SFA), doing data governance and transformation.
A previous student last year (2023) of this module, Alice will talk about how she applies visualization in her work and also about her project last year, for which she won best presentation and poster.
Find 2 bad visualizations and critique them using what we've learnt. Also suggest ways to correct them.
Defining how visualizations are composed of marks and channels. Analyzing the effectiveness of visual channels through accuracy, discriminability, salience, separability.
Workshop on visualization design and validation. Introduction to how the web draws graphics, and the SVG specification.
Color: perception, specification and use. How the eye sees color. Colorspaces. Implications for color use as a visualization channel.
Introduction to D3 (Data Driven Documents), the defacto data visualization library for the web. Toolchains and other libraries. Selections, domains, ranges, axes. Enter, update, exit - design pattern for glassy transitions.
Frame a hypothesis (why), query HDB's transaction data (what), and then design and draw one or more charts (how).
Presentations by students on assignment 2.
D3's geo projection library. GeoJSON and TopoJSON formats. Choropleth viz (D3).
A simple Singapore choropleth viz mashing up population with sub-district data.
A deeper dive into various different visualizations that are used for tabular and hierarchical data. Bar, line, scatterplot, streams, etc. Packed circles, treemaps, dendrograms, etc. Use case examples.
Hierarchical / network chart types. Force directed networks, treemaps, sankeys and arc diagrams. Use case examples.
Charlene is a digital graphics journalist at The Straits Times, where she is involved in producing data-driven graphics and interactive stories.
Her work mostly revolves around topics such as housing and social issues, and has won awards at International media competitions such as WAN-IFRA’s Digital Media Awards. In 2023, she also completed her master’s degree in Data Viz at the University of Girona.
She will talk about data visualization and its relevance to the newsroom.
Assignment 5 is preparation work for the data visualization project, and meant to be presented before the final presentations.
A deeper dive into various different geospatial visualizations. Choropleth maps, isochrone / contour maps, etc.
Leaflet, 2.5D and 3D geospatial viz. Building a simple viz layer on top of a raster base map.
Presentations by students on assignment 3.
We will also do a class-based discussion on some visualization case studies.
Techniques to consider for improving visualizations. Manipulating views - changes, selection and navigation. Faceting and layers. Focus + context (embedding).
David is the head of engineering at AI Singapore.
David is going to talk about SEA-LION, the S$70 million regional open source LLM initiative spearheaded by AI Singapore.
Presentations by students on Assignment 5, which is on the project itself. What is the dataset, why are they doing the visualization (bottom-up or top-down) and how do they intend to put it together?
We'll take a look at a report on the state of the data visualization industry from the Data Visualization Society (DVS) survey 2023.
Guidelines on 3D use, animation and interaction. Latency and feedback. A summary and future perspectives.
Project Development. Consultation by request.
Final presentations.