Visual Vocabulary Discussion Session

This was a presentation and discussion session, covering visual vocabulary (deciding what graphics to use with your data), applications to big data, graphics tools and infographics, with lots of examples. Harriet and Polly ran the session with lots of interaction from the audience (Thank you!). We have attached the slides which were presented here: VisualVocab_Feb25th-20ktc47

There was additional discussion on actually designing your visualisation. Andy Kirk’s webinar – Mastering the Art of Data Visualisation Design – was recommended, as were the books, courses and resources from Edward Tufte. We will look into getting a seminar on the subject if we can find a suitable speaker – recommendations please!

We also discussed infographics and I was particularly interested to hear that some journals do encourage this sort of graphic – indeed, the BMJ actually has a member of staff to design infographics and data visualisations, Will Stahl-Timmons, and we found a good example of his work in a blog post.

The following sites were also recommended – leaning towards the data journalism side with some neat examples:

https://pudding.cool/  “explains ideas debated in culture with visual essays. By wielding original datasets, primary research, and interactivity, we try to thoroughly explore complex topics.”

http://thedataface.com/ “We help brands turn data into interactive stories and analytics tools”

Thanks to all who came to the session and I hope that it was of interest to you. As always, we would love your feedback and please do let us know if you have ideas for sessions you would like to see run (dvg-organisers@sympa.bristol.ac.uk). We really want these sessions to be useful for the group and can only do that with your input too!

12th December – “The Clifton Suspension Bridge Dashboard, Data Visualisation from the Sublime to the Ridiculous”

This meeting featured a talk from Sam Gunner: ‘The Clifton Suspension Bridge Dashboard, Data Visualisation from the Sublime to the Ridiculous’.

31st October – Research data set visualisation ideas – JGI seedcorn competition funding ideas and speed networking session.

Nat presented the Jean Golding Institute’s latest round of seed corn funding, and facilitated a team-building workshop for data visualisation applications to the scheme.

13th September 2018

24th May 2018 – Atmospheric quality data visualisation lab (1)

Guy Barkley from Atmotech introduced us to the work they do and the time-series data he kindly provided for the group on atmospheric quality.

Atmotech is an air quality services business. They deploy IoT sensor modules around sites and present the data to clients – ‘building a picture of air quality’ – in addition to professional services to help improve air quality on site and ultimately reduce exposure to our clients’ staff. They are interested in novel and intuitive methods of displaying our data, which readily engage people and facilitate understanding.

We asked the group to use this data to create data visualisations (as teams or individuals), which will then be presented in a follow-up session to skill-share and learn.

We held some data visualisation labs in 2018, these were two part sessions. In the first part the datasets were introduced and in the second any resulting visualisations were presented. The labs were to work as follows:

• One or two datasets, which lend themselves well for visualisation, will be introduced to the working group.
• Over a time-course of approximately three weeks, members of the working group then create a visualisation using one of the datasets.
• Visualisations can be made using any tool
• In a lunch meeting, the resulting visualisations will be presented and discussed in a round table format.
• There is no prize, the exercise is purely to learn from each other and get inspired. As such, don’t think you have to spend hours on this, just some rough ideas/concepts/prototypes are enough to get the discussion going.

13th February 2018 – Data Visualisation Lab Part 1: Data introduction meeting

The datasets were introduced:

Metabolic reactions of a small human pathogen – Oliver Chalkley

Living cells use enzymes to catalyse reactions to create desired products. Many of these reactions can be connected together to create metabolic pathways. These pathways can be connected together to form a complex network of interactions called a metabolism. The metabolism is central to all life on Earth and can switch between purposes creating a challenging dataanalysis/visualisation problem.

I have included 5 wild-type (e.g. default) simulations of the full metabolism of a small human pathogen, Mycoplasma genitalium. Each simulation contains a time series for all of the 645 reactions. Where the value is the flux through the reaction at time, t. The flux of a reaction is related to the speed of the production/consumption of the molecules in the given reaction i.e. large flux means fast reaction and zero flux means no reaction.

In order to avoid getting bogged down in biological details, I have not included information on the reactions nor the matrix that describes what reactions feed what reactions. This information is available on request but I suggest we see if we can figure out the important relationships from the data.

I have two, related, questions:

1. Is there a good general visualisation(s) to inspect this metabolic data?
2. Using the visualisation(s) from 1. (or some other visualisation(s)) can we see phases within the cell cycle i.e. are certain pathways activated at certain times in the cells life?

NOTE: I have created a software suite for whole-cell modelling (where this data is taken from) and I will try to publish it as a “tools” paper in the next 6 months. Should anyone find something worth adding to the software suite then there is potential to be included in the paper.

Eye movements when reading news headlines – Bobby Stuiifzand

This dataset contains the eye movement data of 19 participants engaged in an information search experiment involving two tasks. In each task, the participant was presented with a set of 10 news headlines and the participant was instructed to either: find a pre-specified word in the headlines (task 1), or, select the headline which they found most interesting (task 2). Each task occurred 50 times (the tasks occurred in a randomised sequence), with the experiment therefore totalling 100 trials, on 50 unique sets of headlines (the headlines were repeated for each task). The dataset contains 84566 rows. Each row in the dataset contains a single fixation-saccade sequence (i.e. “event”), with information on event timestamps, fixation and saccade location coordinates, saccade velocity, and saccade amplitude available. Further, for each row there are numerical identifiers for the trial, experimental block, type of task, set of news headlines used, and participant (anonymised) available.

Interesting questions that can be asked from this data and lend themselves well for data visualisation focus on the difference between tasks, participants, etc. e.g.:

• Do eye movement patterns differ between tasks?
• How do these differences unfold over time?
• Do different participants display different eye movements patterns?

Anyone that decides to work on the data can do so either on their own or in teams, and with their own tool of choice. As we (Oli and Bobby) are the respective owners of these datasets, we are happy to provide some technical support throughout, or to lead on a team working on this data.

6th March 2018 – Data Visualisation Lab Part 2: Presentations and discussion

In this session the resulting visualisations from the data visualisation lab were presented and discussed.