Workshops
From Gephi:Wiki
Workshops are the best way to teach how to use Gephi. Although everyone is encouraged to organize workshops of its own style and content, we provide here some recommendations based on our experience.
Contents |
Organization
Some requirements:
- students should bring their laptop with Gephi installed
- Gephi should also be updated (menu Help > Check for Updates)
- they should also bring a mouse with a wheel.
If you have only one hour, keep focused on the main features of Gephi and answer questions as soon as people ask them.
If you have more time, you could also include a training on datasets. This dataset could be brought by you so you can easily guide your students, or brought by your audience because they want to go head first into their problems.
You could also target a workshop for a specific audience: "Gephi for scientometrics", "Gephi for social network analysis" or "Gephi for semantic networks". This will ease the advertising and attract people naturally involved in the same field, so they could help each others.
Content
Teach the tool
Newbies coming to a workshop should leave with the practical knowledge to use the basic functionalities of Gephi, otherwise this is not a Gephi workshop. Simply follow the Quick Start Tutorial. We advise to show the students how you use Gephi using the scenario provided by this tutorial: they look at your screen and you explain, step by step, why you do each action. Usually the students will also repeat your actions on their laptop, so take your time (~45 min).
Feel free to mix up the content of other tutorials if you have more time, but don't "overflow" them with too much information! In general, we include some tricks of the Visualization Tutorial and provide a quick overview of the Data Laboratory and its Search feature.
Teach the theory
Gephi is designed for people with minimum knowledge on graph theory, however depending on the circumstances you may be faced to people that have almost never heard of networks. You should adapt the content to include as necessary an introduction to:
- graphs (what is a node, and edge, a directed graph, etc.)
- network metrics (what is a node degree, a shortest path, various measures of centrality, basics of community detection, etc.)
- layouts (what is a graph projection, what is the most employed family called "force-directed", etc.)
You might get inspired by excellent courses on networks like Lada Adamic's "Networks: Theory and Applications" course.
Teach the practical
People have data at the end. Although it would be impossible to cover all the way data are encoded and the tools to merge, split, refine and clean them, you can talk about what you usually do, the tools you know, and provide information on the available file formats and databases Gephi can deal with. You may show how to create a network from scratch.
It is a good habit to ask some questions to your audience about their data (nature, encoding) one or two weeks before the workshop. You can even propose them to send you their data so you can check how to put them in Gephi, and let them work on it during the workshop (and not after). They will thank you for that!
Go beyond the expectations
Remember that the individuals interested in Gephi are coming with very various backgrounds and technical skills! It is critical to understand their expectations before starting the workshop to adapt the content. If you can't freely discuss with them before, it is good to ask each one (if possible) to quickly present themselves, the reasons they come, what they know about networks, and what they want to learn. A questionnaire could be given by email a week before the workshop. You may talk about too many things, so be focused on what matters to your audience!
At the end, you may ask them what they liked and what could be improved. If you teach Gephi during a course at the university, students could write a few words on a paper anonymously. But don't bother people by email after the workshop (e.g. difference between Ranking and Partition). Also get noticed on what people get hard to understand during the workshop. We can dramatically increase the quality of our workshops using feedback.
Your audience may also use Gephi in a specific technical environment, and will have questions in it. Here are some often related topics :
- (graph) databases like Neo4j.
- visualization on the Web (Flash, SVG, WebGL...). See a short review here.
