Project Blog

Progress Post Type: ValueAdd
Jisc Tags: disruptiveInnovation;progressPosts;rapidInnovation;VRERI;JISC;BlogMyData
It's well known that creating animations of scientific data creates compelling visualizations that can greatly aid understanding and lead to discoveries. Our first prototypes of the BlogMyData system revealed through user feedback (see First user feedback) that the capability to create animations, annotate them and share them with colleagues, could be a real "win" for the project. Therefore we've spent more time than we originally anticipated on streamlining the whole process of dealing with animations in the system.

The Godiva2 system (the visualization front-end) natively generates animations as animated GIFs, but these have many shortcomings, notably a lack of ability to control the speed of the animation in the browser: a frame-by-frame advance/rewind capability would be very important to users (see Frame-by-frame animations on the web?). We have been experimenting with many possibilities, including producing AVIs for download and using the latest HTML5 features in modern browsers. But this is still under active investigation - if we can crack this problem and make a really nice, streamlined environment for users to quickly create and annotate animations it will be an excellent result for the project. Variability between browsers is currently a major stumbling block- perhaps we are a year or two ahead of time in our ambitions.

Note that we can't tolerate any lossiness in compression, which rules out many animation formats, e.g. MPEG. Producing Flash videos is another possibility, but many Flash movie players are designed for streamed movies, not frame-by-frame scientific videos. (Also, Flash will not work on devices we're keen to target, notably the iPad.)

Progress Post Type: endUser
Jisc Tags: userCase;rapidInnovation;VRERI;JISC;BlogMyData
The project team (thanks Andy and Adit) have got a first end-to-end prototype of the BlogMyData system up and running, with basic capabilities for the user to log in and to blog about map images on the Godiva2 site. So I visited Dan Hodson of NCAS-Climate to demonstrate the system and elicit some feedback.

I went prepared, having met Dan before and worked out which dataset would be a good candidate to show off the new system. We chose a set of monthly means of temperature, salinity and density (key parameters in ocean dynamics) from HiGEM, a state-of-the-art high-resolution climate model. With this dataset loaded up on our test site (having done some extra technical work to correct some metadata headers in the data files), we chatted informally about how the system works, and how it ought to work from Dan's point of view.

Here are the key points of Dan's feedback:

• Dan found it very useful to be able to explore the HiGEM data in the interactive way that Godiva2 allows. For him, it's among the easiest ways to visually get to grips with what's going on with the data.


• But this project is really all about the blogging and collaborative aspects, in which we have a classic chicken-and-egg situation: it's only useful once there's enough information on the blog to make it worthwhile. Therefore we discussed targeting the system to a particular collaboration between Dan and colleagues at the Met Office: the capability to discuss the model over the web could be very useful to help distributed colleagues work together asynchronously. This specific collaboration will hopefully seed the blog with useful information that others can discover.


• The privacy controls are welcome and important: without these scientists would be very reluctant to share their thoughts.


• Although it's nice to be able to blog about static map images, Dan would find it very useful to be able to create animations using Godiva2 and record them in blog posts (and comment about them of course). The development of this capability is now being pushed up the priority list.


• Godiva2's ability to display data in polar coordinates (and to animate in these coordinates) is particularly useful, and timely given Dan's team's upcoming focus on the Arctic. Therefore we need to make sure that polar coordinates are handled correctly within the whole system; this can be tricky, since most "geo-webby" tools use latitude-longitude coordinates, which are not very suitable in high latitudes.

Finally, here's a screenshot of a sample blog entry we generated during the session showing some HiGEM data near the poles:

Progress Post Type: methodology
Jisc Tags: implementation;productivity;rapidInnovation;VRERI;JISC;BlogMyData
Just a few notes on how we are collaborating as a project team. In addition to this public blog (which contains announcements and progress posts; the latter are required by our funders to monitor progress), there are two private blogs, one containing our to-do list and another containing helpful hints we'd like to remember for the future. Although there may be better tools for issue tracking, by keeping all project information on one site we make it easier to collaborate and record information.

As the project manager, I find the to-do blog extremely useful for keeping track of what's going on. This blog engine contains some useful features such as key-value pair tags and Sections, which make it easier to categorize posts and get close to the functionality of dedicated issue-trackers. Unfortunately the unusual tagging system of this blog engine makes it hard to follow JISC's requirements of adding standard tags to progress posts. I hope that having a separate section for progress posts, and a dedicated tag for "Progress post type" will help our funders find the information they need to monitor us.

(The two private blogs may one day become public - but at the moment there is little of interest to outsiders and we don't want to increase the noise.)

Each post in this blog is automatically tweeted, for those of you who have joined the Twitterati and are interested in following our progress.

We use Skype to hold somewhat-irregular meetings of the project team, working off the outstanding entries in the to-do blog.

We are in the early stages of working with users to improve the system; these users are based at the University of Reading, and so I am visiting them personally as often as we can arrange to gradually iterate towards a useful solution.

Jisc Tags: technicalDevelopment;rapidInnovation;VRERI;JISC;BlogMyData
Progress Post Type: techStandards
BlogMyData is all about linking the Godiva2 data visualization website with a blogging engine, so that users can record comments and collaborate on what they see in Godiva2. This post briefly describes the main components of the system and how they fit together.

Here is an overview diagram of the system:


Godiva2 has two extra features: the ability for a user to log in (using OpenID), and the addition of a "Create blog post" link. Upon clicking this, the user is presented with a pop-up window allowing them to add the text and title of their post. This is then sent to the blog engine (in Southampton), along with other automatically-included metadata such as the username, the dataset under investigation and the geographic coordinates of the current viewing window.

The geographic information is stored in a geospatial database (we are using the popular PostGIS system). Actually, for reasons of current convenience, this database is located in Reading, and entries are sent to it from the blog engine over the web using FeatureServer.

The information can then be displayed on the blog's website, or syndicated to other clients using RSS (for Outlook, Google Reader and other feed readers) and GeoRSS (geo-tagged RSS, for geo-enabled feed readers such as Google Maps).