One of my favourite subjects at school was Graphic Communication, which in those days was techie drawing with some other stuff. For as long as I can remember I’ve been interested in how data is presented, in fact the very first comment I got on this blog was from Tony Hirst after mentioning his Visual Gadgets course un-unit (sic) blog experiment – now there’s a fixed point in time.
Since the closure of JISC RSC Scotland North & East I’ve taken the opportunity to rediscover this interest looking at network analysis tools like NodeXL and visualisation techniques and libraries like d3.js and the Google Visualization API. As part of this period of Edupunking myself I’ve submitted some (un)assessments including the Guardian Tag Explorer, which I got some great feedback on and an invitation to visit and chat to some of the Guardian engineers/journalists.
This work hasn’t gone unnoticed at JISC CETIS/OER Programme team and I’m very pleased to announce that I’ve been contracted to visualise the OER Programme (gulp) and I’ve got 40 days to do it (gulp).
This post is day 1 of the project and what I wanted to do was introduce what it is I’ll be doing over the next 39 days, jot down some notes of things I’m thinking of or have already found (including some web standards) and give you an opportunity to comment on any/all of this.
OER Programmes Visualisation Project Outline
The UK Open Educational Resources initiative is jointly led by JISC and the Higher Education Academy and now is in it’s third year. You can find out more about the programme on JISC or HEA pages. I would also highly recommend John Robertson’s (CETIS) collection of posts on OER which are a great lead-in to this project.
So what is ‘this project’ about. Here’s a link to an online version the full OER Programmes Visualisation Project proposal. In brief two work packages, one for examples and workflows for visualising OER project data stored in the JISC CETIS PROD database, and the second mainly to produce visualisations around OER content and collections (these include geographic mappings, subject groupings and volumes).
Initial thoughts on PROD
Having dabbled with PROD before I’m familiar with the data and architecture. Something I regularly struggle with is using SPARQL to query the linked datasets and I usually end up asking Wilbert Kraan (CETIS) for help. Something I thinking might be useful, even if it's just form me, is to create a synchronised version of the data stored in Google Spreadsheets (À la Guardian Datastore). My thoughts behind this are spreadsheets are more commonly understood, Google Spreadsheets include a number of visualisation options which can be embedded elsewhere, and as in my TAGSExplorer the data can be read and visualised using other visualisation libraries. CETIS also already have examples for getting PROD into Google Spreadsheets which means I can spend more time on mashing up the data.
Initial thoughts on OER content and collections
Top level data like geographic mappings look straight forward, but delving deeper is going to take some more work. As the project proposal highlights whilst projects were required host resources anywhere only requiring a reference to be submitted to Jorum what has been submitted (if anything) varies greatly.
OER creation is also only part of the story, after all open educational resources is about providing material for reuse/modification. Where is this data stored. This and other points are highlighted in Lorna Campbell's (CETIS) UKOER 3 Technical Reflections post and these issues also feature in Amber Thomas’ (JISC) Digital Infrastructure to Support Open Content for Education post which supports the JISC OER Rapid Innovation Call for Proposals. [Whilst on activity data there’s a new JISC resources which draws on the 9 Activity Data Programme projects (this includes a section on visualisation) I also can’t but help link to this video which visualises contributions (commits) to Python]
What I’ve learned over recent weeks is VML (Vector Markup Language) works for Internet Explorer 5 and later. It was never passed as a W3C standard, instead they opted for SVG (Scalable Vector Graphics) - Microsoft however didn’t incorporate it until IE9. Some libraries/tools like Google Chart API embed scalable graphics using iFrames and then detect the viewer's browser to deliver the graphic in VML or SVG. One of my preferred libraries, d3.js, only renders in SVG so there is a question about whether it should be used.
I’m only just getting up to speed with HTML5 canvas. Being part of the HTML5 spec it’s only viewable on newer browsers. So like VML/SVG there is a question of compatibility. As canvas converts coded shapes into bitmaps it can render large numbers of objects more efficiently, where as SVG draws the bitmap and remembers the shape allowing further interaction – that’s my understanding anyway.
Top priority is to find out how much UKOER data there is and in what format. Once I’ve planted the seeds of what I need I can get on with the PROD package and the top level OER information. I’ll also need some answers around standards. My feeling is SVG should be okay as it’s a W3C recognised format, but I’m probably biased as it means I can remash some of my d3.js work.
I am as always on the scrounge for new ideas/comments so please share with what’s on your mind ;)