So, when faced with the task of building the Research Object Builder, a couple of goals were expressed: It should be designed to interact with other systems as much as possible, that is, it should take in RDF from some source, decide on what’s useful out of all of that, and then output the useful stuff, after linking it and inflating it so that all commonly used ontologies/tags used to describe an attribute have been added. For sources of data that don’t provide RDF, we should be able to construct shims that use the API of the data source, or other methods, that will output RDF based on the data; these shims can be then fed into the RO Builder to construct or maintain a Research object.
The research object can be used in a number of applications, from archiving, to automatic generation of web pages, and visualization of the relationships between people and research data. The system will thus resemble this :
The design of the RO Builder will allow us to take input from shims that have been developed by 3rd parties, and means that we should be able to add anything to our research object which provides a suitable RDF out, or there exists a shim for generating RDF for it.
Recently I’ve been looking at methods to get all the feeds from the University web presence and compile them all into a big list for later use. This is harder than it looks since the University website is a large beast, with hundreds of sub-domains.
I’ve been writing a basic web spider just using basic command line tools and bash, focusing around wget’s -r flag which downloads all the files in a given domain. It iteratively goes through each subdomain it finds and gets these files too. After each subdomain it deletes all files over 1mb to save space, since it’s unlikely that any web pages will be this large and that’s all we’re concerned about. Afterwards it looks for rss tags in all the remaining files and gives the paths to them.
Currently it’s going through downloading everything, which I imagine will take a long while. Oh well.
To find out what students (and staff) wanted from the new learning environment, we drafted a short survey which asked participants to rank several potential widgets (similar to those on the iGoogle page) in order of preference.
Before producing a fully fledged survey however, we reviewed several of the many online survey tools freely available on the internet (you can read the review here). The two strongest candidates from this were SurveyGizmo and iSurvey. Both were very easy to use, as administrator as well as participant, and generated very professional reports. In the end, we chose SurveyGizmo, as the ranking questions in iSurvey are still in beta and this was the key element to the survey.
The survey also collects a small amount of demographic information such as faculty and year of study, so we can be sure the results are a reflection of everyone’s needs rather than a select few.The main body of the survey is made up of an interactive panel, allowing participants to drag-and-drop responses into the order of their choosing – several participants commented on how novel and engaging this method was.
After only a day, we had enough results to take a snapshot of the data (strengthening our view that students are eager to share their opinions on how the SLE could be developed and improved) and generate a report. The survey will of course stay open, so that more data can be collected to give a larger overall sample size for more representative results.
From this chart the breakdown of faculties can be seen and it is clear that the Faculty of Physical and Applied Sciences (containing the schools of Physics and Electronics and Computer Science) makes up the majority. The is perhaps because these respondents are more comfortable taking online surveys (or just check their computers more often!) – whatever the case, this majority of (assumedly) technical users must be taken into account while the initial results are analysed.
The next question asked the participant’s year of study – in this chart as well there is a clear majority (of second years in this case). However, this is not as much of an imbalance as if there was, say, a majority of staff or PhD respondents.
Above is a subset of the top ranked widgets that people want to see in the new SLE. As can be seen, the top four (so far) are a teaching timetable, linked with a map and a method of quickly viewing the status of outstanding courseworks and submitting them electronically. This gives us an excellent starting point to begin developing tools that users of the SLE will find helpful.
If you are a member of the University of Southampton and want a say in how the SLE is developed you can take the survey here.
What makes for a good RSS feed? Or more specifically, what makes an RSS feed look attractive when viewed in a ‘Personalised Magazine’ format. That’s what I’ve been trying to find out and have come up with three (plus one control case) potential improvements. These all utilise the ECS ePrints RSS feed which can be accessed here: http://eprints.ecs.soton.ac.uk/cgi/latest_tool?output=RSS2.
This is the base case consisting of the default ePrints RSS feed. Not much to say about this one: the body consists of just the authors and there are no embedded images.
Case 1: PDF Preview and Article Abstract
In Case 1, the ePrints generated PDF preview image is used and embedded in the RSS item. Also, I have replaced the default list of authors with the abstract of the paper which is much more useful.
+ Easy to find and insert preview for items where one is present.
- Preview not present for all items in feed.
- In many cases is not much more interesting than the control case.
Case 2: Article Abstract and Image Extracted from PDF Document
In this case I have again used the abstract as the body, but for the image I have used an image extraction tool to extract images from each PDF, and then hand-picked the most interesting one from each item, and embedded that.
+ Relevant to the item.
+ Can sometimes give a more attractive image than case 1.
- Not all items have a PDF present, and not all PDFs present contain extractable images.
- Many items do not contain attractive or interesting images.
- Hard to select which image is the best automatically.
Case 3: Article Abstract and Keyword Image Searching
In this case I again used the abstract, and also added an image which was found using keyword search for creative commons images on Flickr. The images were sorted by “interesting”, and the top image selected. The keywords were selected by hand from primarily the keyword field if present, or if not by hand selecting an important word from the title.
+ Images can be found for every item.
+ Images are more exciting and interesting in most cases.
- Hard to select keywords, especially in cases where keyword field not present.
- Images are the least relevant to the item of all the cases.
- Some selected images may be unsuitable for purpose.
The first part of any large project is finding out what your client actually wants. We want the SLE to be not only functional, but enjoyable to use as well and, with this goal in mind, want to include the student body as much as possible during development.
Focus groups have been an integral part of early planning, but to reach a wider range of opinions online surveys are invaluable. There are many options of survey tools available on the internet, of varying price and quality. A selection of the free services are looked at below:
This site is very popular, and easy to use – when you create an account it is possible to take a short tutorial in the form of a questionnaire to show how the final product could look.
While containing a wealth of features when creating surveys (such as randomizing questions to remove order bias) it is let down by many of these being restricted until a full account is bought.
Also an easy site to use withe a wide array of online support, and has the added benefit of offering prebuilt templates on top of the option to build surveys from scratch.
Various themes can be applied, including several that are smartphone friendly.
Reports are particularly professional, showing not only a table of responses but also demographic information such as which country and city the respondents are from.
Zoomerang is a very business orientated tool and (possibly as a result of this) has a heavily restricted free version. It is not as intuitive to navigate as any of the other sites – however, the questionnaires themselves are visually appealing, such as interactive drag-and-drop ranking.
Zoomerang also offers prebuilt templates, with industry specific questions.
The results are a little difficult to read in the provided table format, and there are very few options for analysis available in the basic tool.
Although some users may not realise it, Google Docs also allows its users to create and publish simple surveys (under the “Forms” category).
Though not as popular as the sites looked at so far, it nevertheless provides a very functional service, with the added bonus of being able to integrate the results directly into a Google Docs spreadsheet.
It is let down by its “bare bones” style – not being a dedicated survey tool, it does not offer all of the functionality as other sites such as ranking questions or demographic information.
If you only want to ask one question to a large (potentially huge) number of people, it is worthwhile considering social networking. Facebook, for example, provides a “Poll” function which gives the user a simple interface to ask a question and post it to their wall. Friends can then either respond with the given options or (if the asker permits) add their own.
Once they have responded, the poll is added to their wall and so on – in this manner popular polls can reach a staggering amount of people in a short amount of time. However, the poll may not be relevant to all of the people who respond (as in the case of the SLE) and the results are publicly visible which can cloud the accuracy, making it unsuitable for a scientific survey, but perfect for a quick survey of popular opinion.
The University of Southampton provides its own survey tool to staff and students, iSurvey. This stands up surprisingly well in comparison to the “big-name brands”, offering wide customization (right down to the labels of the “Next” and “Previous” buttons).
Like SurveyMonkey, SurveyGizmo and Zoomerang it offers many different question styles and in addition has the added benefit of knowing exactly where the data is stored. The obvious downside is that it is only open to staff and students of the university.
There are various options open to result analysis including charts, tables and statistics.
I will use the term Personalised Magazine to describe the social media and news aggregation applications I discussed in the last post. It seems to fit the purpose nicely and makes sense.
There are lots of different apps and many seem to offer something different, be it platform or special features. Some are free and some are pay only. Here’s I’ll briefly go over 3 free personalised magazine apps. I’ll use the example of the ECS Eprints RSS feed to show how they all present the same data.
Feedly is a multi-platform and multi-browser application. It has apps for both Android and iPhone, and also has plug-ins for Chrome, Firefox and Safari. On the desktop it runs in-browser as a web page. It gets the user to login to their google account, and uses their Google Reader feed to generate a magazine. Feedly is free for all versions.
Opera RSS Reader
This one is a bit of a misnomer since it’s not really an RSS reader at all. When you visit an RSS feed page in Opera, a magazine style preview page is generated for that RSS feed. However Opera won’t actually let you use this to view all your feeds in magazine format unfortunately. This seems to be an in demand feature so maybe they will add it soon, but for now it’s not possible to use this for your general viewing purposes. It’s worth taking a look at though to see how it’s presented.
Pulse is a mobile device only app (comes on iPhone, iPad and Android) which takes your news and arranges it in a colourful attractive mosaic. This one doesn’t really look like a magazine but it fulfils the same purpose, and looks nice in the process. This one offers many ways to pick which content you want to look at, including directly entering an RSS feed link, but also connecting to a Google Reader account, and picking from one of many news sources from Pulse’s extensive library.
Most of these share the same problem in that because the ECS news feed isn’t very colourful, the magazine page ends up looking the same. Feedly tried to improve things by adding stock pictures but didn’t have much success. I have used the Engadget RSS feed as an example of an “interesting” news source. That is, one which generates nice looking pages on these apps. The main difference is the presence of an image in the RSS feed itself which is used as a thumbnail or illustration for the article in question.
There’s currently a problem in the way that much research is judged almost exclusively on papers written about it – in the current day there are frequently many other different aspects to the research. The problems associated with this were talked about here, and so, to simplify the task of using a larger set of data to rank research by, and to promote wider use of RDF, my internship will be spent developing a web based, distributed, Research Object creator, that will allow easy creation of Research Objects, which will function as a description of a piece of research, and as an aggregate of useful data related to that research.
Currently it is difficult to keep up with research progress and developments even within your own University campus. News is usually divided into lots of separate RSS or Atom feeds which do not tend to present their information in a very inspiring or exciting manner. Thanks to the JISC funded CampusROAR project over the course of the summer I will be investigating methods to present the information published by department and/or university research news feeds. Hopefully our representations will be more colourful and interesting and engage a wider audience utilising or taking inspiration from applications such as Flipboard, Pulse, Zite and other magazine-style news aggregators.
The JISC funded CampusROAR project has several strands of work. Keep an eye out for our logo as you are browsing around the web.
My internship consists of helping develop the new Southampton Learning Environment (SLE) – “a 24/7 ‘one stop shop’ for learning and living needs at the University of Southampton” as the research team’s website puts it. When finished, it will include various “apps” such as email, news feeds and more, as chosen by the student body.
One element of my part in this project entails gathering student opinions on the current environment (“SUSSED”) and how they think it could be improved.
This will be supplemented by finding data stores that would be useful to staff and students (such as bus time tables or swimming pool booking databases) and converting them into a machine readable format (if they are not already), so they can be integrated with the SLE.
Associated with this will be drawing mockups in Pencil (an open-source 2D animation software) to give an idea of the way the environment will look. These mockups will then be used to create a more realistic looking (but semi-functional) HTML mockup of the final product.