Visual Analytics

Galas initial thoughts on Visual Analytics.

• Visual analytics and leaning analytics can guise itself for presenting itself as learner/user centered but in reality the real goal may be to collect data and findings for individual researchers, or the wider institution.

• There is space for the research of Critical Realism (and the philosophy of perception) within this field (gap?)

• In terms of values, how is the individual treated? Permission must be sought – in this information age we must still consider privacy and the rights of the individual.

• Who is the data really for and what is it being used for? Need for transparency.

• The software development for visual analytics is determined on the data and purpose. This appears largely to be a case-by-case basis due to a variety of possible data sources – a universal set of principals can be established (as they have been) but contextually relevant design require expertise development. (time and money, special area of expertise)

• How much visual analysis in actually in action? – Funding and allocation towards this specialty?

• What can we learn if anything from the visual analysis of learning analytics from MOOCs?

• The so-called need for visual analytics comes from the so-called influence and abundance of the impact of educational technology in the classroom and online environments, and the so-called ‘displacement of the teacher’ within this. Critically speaking, these are all highly debatable claims and areas within higher education which have various truths and un-truths depending. (The arguments are exactly that and not necessarily reality) Thus the myths and determinism around ‘data’ and technology within education and empiricism are driving the VA Agenda. Thus it is important to note certain assumptions authors have and the field generally holds. So that we can determine the hype and the ideology from actual impact etc.

• There is a leaning towards empirical findings from data. This is common but suggests that there is space for qualitative approaches to VA such as interviews with teachers etc.

• Is this all a reaction to the mechanization of education  Do teachers now not have enough time to talk to students one on one or are they just trying to justify having a computer do it for them..etc? is dealing with data in this way closing the loop or just getting further from the truth/at-ness?

• Should teachers play the role of ‘visual analytic tool’ are their own observation and their own brain and intuition and common sense and gut feelings and experiences being disregarded within all this or merely supported?

• VA is not ‘new’ simulations modeling has been happening for a while – a timeline of technologies and systems could help here.

• Where does the judgment lie? WHEN are decisions made?

• Ontology – who has the eyes – man or machine?

• Any software with user interaction often require training or expertise, you can’t just say you don’t want that.

• How does time and speed come into this? Is the VA movement just another tool to try to quicken efficiency? Is VA then de-humanising?

• Modernist design and utopian universalism is inherently problematic – what of personalisation, culture, accessibility ete. All users are different. – the decision of ‘organisng principals’ will be determined on the context, data and task at hand.

• In terms of the above – modular/parts thinking design such as the ethos of ‘prefuse’ may offer a solution. Open source and customisable and re-usable. Seems to be a sustainable  flexible and collaborative way forward.

• Chaos theory, AI and fractal logic could play a role in prediction.

• We can borrow knowledge from systems thinking and modeling  (such as boundary making and the purpose and context triangle)

• Can we create a domain specific infoVis application for Higher Education?

• Even if toolkits or software becomes usable and effective – does it make it ‘worth while’?

• In terms of the design ethic ‘form follows function’ we need to arise and adhere to this. We must be careful not to fit data into pre-established methods when we have a chance to radically alter the way we organise and perceive data in the first place. The medium is the message (McCluchan) and John Maeda (self referencing technology and copy cat transmissions). (his rules of simplicity) Does big data now have a new function in terms of information network etc? more discovery approach than representational – a creative space? A constructive space? A playful space? 

• Theory holds high importance for VA, we must remain critical and well thought out – don’t want idealism etc.

• In terms of language and visual language – what is going on here exactly? How much is mental modeling  What can we learn from nature – morphology and our own natural ways of decoding information (pre-established and ‘new’) adding an extra dimension doesn’t automatically make it new etc.

• Pre-established knowledge in how to read graph types is necessary – this isn’t necessarily as ‘intuitive’ as claimed.

• Where are we aiming the usability and and what skill level and skill sets? Who is the audience? We can’t just assume ‘one universal audience’ it could go anywhere from a computer illiterate to a highly specialised programmer.

• Designers have been successfully doing InfoViz for years – creating and playing in their own software.

• My own custom area of speicalisation could be in ‘rendering’ .. custom action types etc – this could get pretty creative!

• Need more testing for the end user being a lamen – getting them ‘thinking aloud’ ie. teachers.

• Issue of raw data coding and manipulation – authors neglect the human component in this – this is where a lot of the analysis even begins!

• ‘Information Aesthtics’ seems my particular space, another claiming a 'new' term.

• the shift is a democratization of visualisation. This shift co-incides with a shift in design and technology towards distributed networks.

• What is the “raison d’etre” (reason for being or existence) for data in the context?

• The number of approaches towards data visualisation highlights the importance of design and creativity in the visualisation field

• If it’s about engaging ‘lay’ people in understanding complex data insights – shall we not just get them to play with lego?

• Little research is on the design process of visualisations.

• The semiotic taxonomy is a step in the right direction – along with pre established methods – well thought out theory and an open interdisciplinary approach we need to also build a library of key terminology and skill sets.

• Nvivo and such data analysis software deals with this sort data (dynamic and diverse) for example and we could learn from them.

• VA can’t ‘make meaning’ for us – just visualise data. There still needs to be someone doing the work of coding all the bloody complex data! Is this not the biggest issue? Are we not sidestepping this issue?

• Dromology and time space compressions of post modernity seem relevant here.

• Considering the VA Agenda – how much of the ideology associated with terrorism etc. enters VA – why – how and to what affect? The classic military and tech link. Also, an emergency response involves time and pressure – so ‘illuminating the path’ reading in particular which a lot of the lit seems to refer to may not be that relevant for other areas.

• The big question seems to be can we apply human judgement to complex data in pressure filled situations? I’m not convinced that our brain can conceive – VA may be a bridge to understanding but not necessarily a solution to understanding big data. We may need trained specialists (think Minority Report)

• Why is new tech and new talent suddenly “urgent” now???

• What type of thinking does it take to be an analyst? Who is an analyst? Is this a specialisation and skill set?

• What do these so called ‘next generation’ of technologies look like? What makes them ‘new’? do we need an overhaul? A whole new paradigm?

• When is data ‘big’?

• Is this more of an AI issue or wanting to build a big growing adaptable mainframe?

• VA often used for ‘reducing risk’ this is interesting and opens up lots of ethical concerns

• Arguments against the science of analytical reasoning?

• Are the VA agenda principals even applicable for HE considering the principals should determine the design

• Key: “To provide task-appropriate interactions that let users have a true discourse with their information”  Here there needs to be a greater emphasis on reflection and play.

• Interaction designers will have a lot of answers to these questions, the design field needs to work collaboratively with academics in other fields and vice-versa. We can learn a lot from interaction design – i.e. interactive exhibits. (And mapping interaction approaches to analytical tasks)

• Does everyone have the same level of visual literacy and understanding of visuals? I think not – this has failed to of been mentioned. Some are more apt at visual thinking and encoding/decoding visual information than others. – This also depends on the level of complexity – but the more abstract you risk either getting further away from the truth or closer to its purest form.. this argument of abstraction is a classic one and people argue either way. It’s often a threshold or point of reference  We need to maintain a maximal to minimal bottom up approach  When dealing with individuals we deal with subjectivist  culture etc. there are qualitative differences to consider in this debate not just quantitative  Semiotics hold a universal solution here > visual metaphor etc > but modernist deisgn has its pitfalls. We need a post modern approach.

• History needs to be referenced more – how is this new? We have ben visually analysing stuff for eternity. Let us not forget.

• Applying physical concepts like mass and gravity help people understand data and its true ‘weight’ (applying physics) as do common notions of space and time and narratives.

• Data types are continually changing – should we just assign everything with binary 0’s 1’s ?

• Whats this fascination with ‘reducing time’ good things take time! Speed and society – Virilio etc.

• Google analytics seems like a fine start point.

• Interoperability needs to be forward thinking – this is super difficult, as is Component-based software development. The computer is becoming organic and self organising..