From Connected Data to Meaningful Visualisations
A Talk by Evmorfia Argyriou and Benjamin Niedermann
About this Talk
Graph visualisations make complex data more accessible, but creating a meaningful graph visualisation is far from trivial. In this masterclass, we present a pipeline for transforming raw data into a visualisation that answers specific user questions.
This class is designed for anyone working in data analysis, including data scientists, application developers, and researchers. We use free tools accessible to everyone.
By the end of the masterclass, participants will have gained hands-on experience in transforming raw data into meaningful graph visualisations.
Description
Techniques of data analysis allow users to extract valuable insights from raw data. By exploring this data, they can answer specific questions, identify patterns, and make informed decisions. Graph visualisations enhance this process by making complex data more accessible.
In this masterclass, you will discover how to transform raw data into meaningful graph visualisations, covering the entire pipeline from data to the final visualisation. First, we provide an overview of how to identify key questions that can be answered with graph visualisations. Then, using Jupyter Notebook—a popular and free tool for data analysts—and Neo4j databases, you will learn how to prepare data and apply graph analysis algorithms such as centrality and other relevant measures.
Finally, you will explore how to select the appropriate visualisation style, focusing on what should be emphasised and how information can be effectively encoded in the visualisation.
Key Topics
- Graph visualisations: enhancing data accessibility and making complex data easier to understand
- Graph analysis techniques: extracting insights from structured data and identifying patterns
- Visualisation style selection: choosing appropriate visualisation styles based on data
- Information encoding: emphasising non-structural data in graph visualisations
Target Audience
- Data Scientists and Analysts
- Data Engineers
- Application and Database Developers
- Researchers and Academics
Goals
Get hands-on experience transforming raw data into meaningful graph visualisations by identifying key questions, preparing and analysing data, and selecting appropriate visualisation styles to effectively communicate insights.
Session outline:
- Part 1 – Brief introduction
- Questions that can be answered with graph visualisations.
- Aspects to take into account when creating a graph visualisation.
- Part 2 – Prepare the data
- Present the dataset and the underlying graph structure.
- Properties of the graph structure: e.g., centrality, or other measures.
- Part 3 – Visualise data
- Choose the most effective visualisation style and determine which elements of the structure should be emphasised.
- Style the graph visualisation: Decide how to encode information through the styling of nodes and edges.
Format
The masterclass combines a lecture format with hands-on coding exercises.
We will be using Jupyter or Google Colab notebooks (participants can choose) with standard Python packages. Data will be retrieved from a Neo4j database, which we will provide. For graph analysis, we will use the Graph Data Science Library. For data visualisation, we will utilise the yFiles Jupyter Graphs for Neo4j widget. This is an open-source wrapper for yFiles Jupyter Graphs that offers a simple way to visualise Cypher queries executed against Neo4j databases.
Level
Beginner - Intermediate
Prerequisite Knowledge
Basic Python and familiarity with standard packages (pandas, numpy, scikit-learn)
We’re a cloud tech company that provides organisations around the world with computing infrastructure and software to help them innovate, unlock efficiencies and become more effective. We also created the world’s first – and only – autonomous database to help organise and secure our customers’ data.
