Month: October 2011

Categories
Mobile User behaviour

Early MoveMe wi-fi heat maps from Love Lace exhibition

Several months ago I announced that the Powerhouse Museum was a partner in the MoveMe pilot project funded under NSW Government’s Collaborative Solutions Program.

We’ve been working with Ramp, MOB Labs, ShopperTrak and Smarttrack RFID to deploy the pilot in our recent Love Lace exhibition.

This exhibition is ideal for trialling location aware content delivery because it is already kitted out with public wi-fi and we have the cross platform iOS and Android free exhibition App. Even better, the exhibition uses QR codes and the QR code reader in the exhibition App which gives the pilot project a great baseline to compare usage against.

While we don’t yet have the location aware content delivery working – that will come in a future version of the exhibition App – we have started to get access to wi-fi tracking data using the ShopperTrak system. As explained by Christopher Ainsley & Julian Bickersteth in their paper for Museums & the Web earlier this year, the ShopperTrak system is already used to create heatmaps and visitor journeys through shopping centres (or ‘malls’ as some readers might describe them).

The first data has started to emerge from the system and it is already very interesting.

Here’s a dwell time heat map that shows the areas of the exhibition where the wi-fi enabled devices (presumably carried by visitors) spend the longest time. This shows data from Sunday Oct 30 and 226 tracked devices.


(click for larger version)

A couple of important caveats.

Whilst the sample sizes are unexpectedly quite high (largely because the wi-fi tracking doesn’t require an actual connection to our wi-fi network, just that it is switched on on the device/phone), the sample rate at which devices are ‘pinged’ is quite low. iOS devices, for example, are only pinged every 2 minutes and so the resolution is very low – unless they are actively connected to our wi-fi network for the exhibition. This means that if an iOS device has wi-fi switched on but they aren’t using our Love Lace App and not connected to the exhibition wi-fi and they spend 10 minutes walking around the gallery their device will be counted in a maximum of 5 locations. Of course this can be offset by the volume of tracked devices (which almost certainly exceeds that of other manual people counting methods employed by traditional audience research).

What is interesting about the data is that it pretty much mirrors the distribution of the QR code usage I blogged about earlier. Unsurprisingly the longer dwell times are where the sit-down video experience is.

Categories
Mobile User experience

Experiencing The O at MONA – a review

A lot has been written about the Museum of Old and New Art and I’m not going to rehash any of that. Instead I’m going to look at their mobile guide – The O – which is provided to every visitor and included in the admission price.

Here some of the fleet of 1300 Os sit charging in enormous custom charging bays where they can also be updated.

The O is an iOS App that runs on an iPod Touch comes ready to run and with a quality pair of Audio Technica headphones. Developed by Art Processors, The O is described thus;

Wall labels are at once didactic and limited. They inhibit imagination. Squinted at through a dozen huddled heads, they are barely useful tools for learning, much less free thinking, or a private appreciation of the objects they describe.

The O solves these problems. It delivers information in a way that enhances the visitor’s experience of the gallery, and enables curators and exhibition designers to display the works the way they want. Museum researchers can present the best, most relevant textual, visual and audio content at their discretion. It provides information on visitor viewing habits, trending and satisfaction via integrated statistical reports. Above all, The O is an intimate, intuitive interface of the learning and autonomous response.

None of this would matter if it was a pain to use.

I was very impressed by the ‘technology concierge’ skills of the ticketing staff – they run you through the basics of the App and the hardware as they sell you your ticket and set you off on your way. Sitting beside the cash register is a graphic clearly explaining each of the main interface screens of the O as well. I’ve never seen this level of ‘scaffolding’ happen in other museums and the deftness with which visitors are set off on their way quickly is a testament to their staff training (and acceptance amongst these staff of the value of the O itself).

Descending into the museum itself you launch the O and you are off. Pop up instructions help you through the basic App operations and after a while you are prompted to enter your email address (and optional country) to ‘save’ your journey to the MONA website. Once this is done there are no further prompts and even when, as I did, returned after lunch and was given a different O device, the final ‘saved tour’ seemed to accurately aggregate my whole visit (over the two different devices).

At its most basic level The O replaces wall labels. Entering a space you simply click ‘Update’ and, using wifi triangulation a proprietary real-time system (see comments), the device provides a list, with thumbnails, of objects ‘near’ you. This works surprisingly well despite the split levels and bulk showcases of coins and other small objects in some areas. The scrollable list relieves the technology of the difficult task of ‘exactly positioning the visitor’ whilst at the same time emphasising the visitor’s own agency in choosing what they are ‘seeing’. (I think this is going to be an increasingly important balance as location and compass headings give mobile devices better granularity at guessing what you are looking at).

However the most impressive part of The O is the content – not the technology.

The O provides simple label text and an image for every object. I was disappointed that the images weren’t zoomable, however on most objects there was also a curator’s piece amusingly titled Art Wank. These were short, very accessible and gave useful context and background without overdoing it. A slightly smaller subset of objects are augmented with options called ‘Ideas’, ‘Gonzo’, and ‘Media’. It is in these three areas that The O really differentiates itself from every other museum mobile App or guide I’ve experienced.

‘Ideas’ is simply a set of provocations – or talking points. Some are quotes, others are just statements. One of the many ‘delighters’ I discovered on The O visiting with my companion (with her own O), was that often there were multiple ‘Ideas’ and that very rarely would we both get the same one at the same time. This gave us prompts to talk to each about the objects we were looking at – ensuring that sociality was not eroded by every visitor being glued to their own screens.

‘Gonzo’ is almost mostly responses or stories from MONA’s owner David Walsh. Sometimes these are stories about the acquisition of various objects, other times they are hilarious, for want of a better word, ‘rants’ about the artist, a style, or a moment. Like the ‘Ideas’ they make great talking points.

‘Media’ are short audio files – interviews with the artist and others. Some objects also have songs by Damien Cowell who was commissioned to record them ‘about’ certain works.

The interviews blew me away.

Unlike every other ‘museum tour’ the audio interviews are completely raw and lo-fi. This shocked me – and I loved it. Almost all the interviews that I listened to sounded like they were recorded in a noisy cafe – and in more than a few the interviewee’s mobile phone rang in the middle of the recording (usually followed by an apology ‘sorry I’m in a meeting’). This made it so approachable and friendly – and, importantly, felt candid – like I was there with the artist. This also reminded me that the quality of the content always trumps the fidelity of the recording.

‘Loving’ or ‘hating’ objects is possible too, and doing so gives you a simple quantitative statistic on the objects popularity amongst other visitors. I did wish that this recommended me other things to go and see. I also missed any kind of search functionality – I understand that this is probably because ‘searching’ is the exact kind of intentionality that MONA is trying to disrupt, instead forcing you to be in the moment – but it was frustrating when there were certain works I knew about that I wanted to locate.

Leaving MONA, the headphones and The O were given back to the friendly staff at the door. Arriving back home, there in my inbox was an email from MONA linking me to their website where I could browse through the objects that I’d seen – after supplying the email address I used to register), and find out which I’d missed.

The post-visit web experience is interesting in that it requires a MONA visit (and user registration through The O) to get access. On one hand this might seem exclusionary – and is definitely an option that is really now only open to private museums with no public mandate – but on the other hand this did re-emphasise the importance of connecting the physical experience of MONA and its works with the online experience. And, that I couldn’t access the objects of the museum before my visit (beyond a few selected pieces), meant that I was more open to exploring than targeting only things I was interested in when I was in the galleries.

On the web you have access to all the same content you could get on The O – the audio, the text, but rather disappointingly only the same small size of image. Your path through MONA is visualised and able to be played back on a timeline. I’m not sure that this adds any navigation ‘value’ but it does re-emphasise the physicality of visiting MONA, its unique spatial construct, and its primacy in understanding and experiencing ‘the works’ inside.

This is one of the few examples of where a museum website actually enhances the post-visit experience by connecting it concretely back to the physical experience (and does so be explicitly preventing pre-visit planning and expectations).

There’s a couple of minor quirks (primitive audio player controls especially) with The O but overall it sets a new benchmark in terms of integrated interpretative devices.

I do wonder, though, how much it relies upon a few uniquely MONA attributes – its entirely private vision (versus public duty/mission), the design of the museum itself which prevents any other form of internet access (it is underground), and the tabula rasa upon which it has been able to construct its content all at once (no legacy material or practices to deal with)?

And, how the aggregate usage data – the loves/hates, the pieces that are most/least viewed, the contours of content – is used will be fascinating to see.

Categories
General

We have a new address – Freshandnew.org!

You might have noticed the web address of Fresh & New(er) has changed. We are now running at http://www.freshandnew.org and hopefully everything, including RSS feeds, is working as expected.

Categories
Conferences and event reports

Culture + heritage + digital at Web Directions South 2011

Sebastian Chan and Luke Dearnley

Luke Dearnley and I were last minute additions to the Web Directions South lineup last week. Coaxed by Maxine Sherrin to do a ‘fireside chat’ we sat comfortably by a digital fire and talked broadly around some of the exciting projects that are happening in the digital heritage space right now.

We tried to cover a lot of ground and tease out some of the issues in the sector as libraries and museums around the world finally begin to build significant momentum around digital content. Taking these discussions to the web developer community is important because all this is happening at a time when the government is calling for discussion of the National Cultural Policy where there is talk about ’emerging technologies’ and the NBN in the ‘arts’. (See the Ideascale on the digital culture response to the NCP.)

Here’s a brief rundown of what we covered in our free-wheeling talk done without notes (and, sadly, much sleep).

I started out looking at where we were at the Powerhouse in 2001. Back then we were talking about the ‘virtual museum’ and exploring 3D tours and building monolithic encyclopaedic resources using our ‘authority’. Whilst there was some amazing stuff built back then, that won awards (and we still get enquiries about), the web has changed.

And now where we are in our thinking in 2011.

Now it is all about being a data provider, getting the our knowledge and collections out into the community where they can be debated and gather feedback and attract interest. The social web and now the mobile web has made this possible at the kind of scale that wasn’t possible in 2001. At the same we now have ‘contextual authority’ rather than what we previously imagined was ‘overall authority’. Remember that in 2001 Wikipedia was only just starting and had only 6,000 articles.

At the same time the user is firmly in control not only of how they navigate ever growing competing information sources, they also are using interfaces that fundamentally change how they perceive their computing devices. Touch and now voice interfaces, radically personalise, even anthropomorphise our devices. They are carried closer to us than ever before, creating a sense of intimacy and helping us form (unhealthy?) relationships with our mobile technologies. (“Excuse me while I just check my iPhone one more time – I haven’t touched it in the last five minutes.”)

In the background of this slide you can see an early heat map that is produced by tracking the dwell time of visitors carrying wifi devices in one of our exhibitions (they don’t even need to be connected to our wifi to be picked up). I’ll be blogging about that shortly in a new post but for now it should serve as a reminder that this sense of personal connectivity comes at a high price of personal trackability. It isn’t simply bundled up under ‘privacy’ and there’s a long way to go in the public discussions and debate about the trade off between utility and privacy.

The other big change is that of scale.

A collection like that of the Powerhouse used to feel ‘large’ but in actual fact it is tiny. It’s value in the digital space now is no longer as an island but only in what it can contribute to national and international collections – a collection of collections. That’s a tough challenge for a State-funded museum whose majority of ‘visitors’ walking in the door live in Sydney.

But at scale new possibilities emerge.

At this point we started to look at some of the initiatives that are exciting us around the world at the moment. Initiatives where the ‘value’ wasn’t necessarily obvious at the beginning but emerged only after time.

We showed and talked about ->

– Tim Sheratt’s work with the digitised newspaper collections in Trove and the emergent stories he is starting to knit together by analysing the changes in language in newspaper articles over time, or by facial recognition in archival collections. These stories are only possible at scale – and even now they are terribly incomplete with uneven digitisation of each State’s newspapers in Trove – but they are getting better over time. Everyone (even you, dear reader) needs to go an read the transcript of Tim’s recent keynote at ANZSI. We are at the very very beginning of this but Tim’s work hints at some of the possibilities.

– New York Public Library’s historical menus project and how marking these menus up in the way they have lets us observe the changes in diet and ingredients, as well as food prices over time. And how, of course, dining at the Possum Club in 1900 would have been quite an experience.

– The other thing about the NYPL menus project is the way that, prior to releasing an API, they’ve done what we did at Powerhouse. They’ve released the whole data set as a ZIP. As we found with our own collection, a downloadable full dataset allows people to do mass scale analysis more quickly and easily (and with less drain on your server) than using an API.

– Looking at scale we briefly showed the free ImagePlot toolkit from the Software Studies Institute at UC San Diego, and how it by allowing you to do image analysis of enormous corpora of image files new patterns and relationships can be discovered.

– Luke talked about linked data and how connecting everything up is slowly becoming possible as more things and thesauri go online. We showed a couple of nice front-end examples of some of the possibilities when collections get connected up. Our very own infant site – the Australian Dress Register – which is slowly growing and bringing on new contributors; and the newly re-designed and re-configured Design and Art Australia Online (formerly Dictionary of Australian Artists Online). Here’s a biographical entry for one of the designers with lots of objects in the Powerhouse collection. Here it becomes possible to traverse her ‘associates’ as well as all the exhibitions etc she has been involved in all over the world.

– We looked at some other exciting community transcription projects that are overcoming difficult issues of both relevance and specialised content. We showed the fantastic Old Weather project with the Citizen Science Alliance using old ship logs from the National Maritime Museum to gather geolocated climate data form the past. It is one of our personal favourites and Fiona Romeo at the NMM published a great paper on it at Museums and the Web earlier in 2011 which you should read. What we find really lovely about this project is that it finds deep value in the kind of collection that museums find very difficult to ‘exhibit’. Actual ships – easy and attractive to put in an exhibition but the ship logs – much harder.

– We also showed the interface for another Citizen Science Alliance project called Ancient Lives. This project is getting citizens to help transcribe papyrus scrolls from the Oxyrhynchus collection whose story of acquisition and discovery is enough to encourage you to give it a go.

In wrapping up we started to ask a number of questions that remain unanswered/unanswerable:

– what the barriers to a Europeana-like project are in Australia, let alone a Digital NZ? Are they more cultural reasons than anything else? What is of ‘national significance’ that we can all agree upon? Is such agreement even possible in a fragmented nation?

– does the ‘open’ in linked open data matter more than just linked data in the short term?

– are libraries able to knuckle down and focus on digitisation better than museums because they aren’t expected to ‘also do exhibitions’? This looped back to an early slide where we talked about the ‘post-web accord’ that emerged in the mid 00s. Is this accord coming under pressure as a result of changing economic circumstances? Or is this just one of the many museum challenges that are under discussion in the sector.

Categories
Collection databases Interviews User behaviour

“Do curators dream of electric collection records?” Exploring how the Powerhouse online collection is used

As one of the first of a ‘new style’ of museum online collections, launching several internet generations ago in 2006, the Powerhouse Museum’s collection database has been undergoing a rethink in recent times. Five years is a very long time on the web and not only has the landscape of online museum collections radically changed, but so to has the way researchers, including curators, use these online collections as part of their own research practices.

Digging through five years of data has revealed a number of key patterns in usage, which when combined with user research paints a very different picture of the value and usefulness of online collections. Susan Cairns, a doctoral candidate at the University of Newcastle, has been working with us to trawl through oodles of data, and interviewing users to help us think about how the next iteration of an online museum collection might need to look like.

I asked Susan a number of questions about what she’s been discovering.

F&N – You’ve been looking over the last few years of data for the Powerhouse’s collection database. Can you tell me about the different types of users you’ve identified?

Based on the Google Analytics, there seem to be four main types of OPAC users. I’ve given each of them a nickname, in order to better identify them.

The first group is the FAMILIARS, composed of people who access the OPAC intentionally. FAMILIARS know of the collection through either experience (having used the online collection previously, or from visiting the museum), or via reputation (ie GLAM professionals, researchers or amateur collectors). FAMILIARS come to OPAC with the highest level of expectations and have the most invested in the experience. Trust and authority are hugely important for the people in this segment.

The second group, I’ve called the SEEKERS. Like FAMILIARS, SEEKERS are driven by a desire for information they can trust. However, unlike FAMILIARS, SEEKERS do not yet know about the museum and/or its collection. This group includes people who are new to collecting communities, or student researchers etc. If they find what they are looking for on the OPAC, SEEKERS have the potential to become FAMILIARS.

The final group for whom authority and trust in information is important are the UTILISERS. These visitors, primarily education users (like school students), have specific and particular research needs, which are externally defined (ie they might be looking for answers to set questions). This group is task-oriented.

The last group that comes to the OPAC is the WANDERERS. These are casual browsers who seek fast and convenient information, but don’t necessarily need depth in their answers. Seb once nicknamed them “pub trivia” users, and that seems pretty apt.


F&N – What sort of proportions do each of these make up?


By far the greatest number of OPAC visitors are WANDERERS. More than 80% of all OPAC users – whether in a two-year period, or a six-month timeframe – visited the collection online once. Obviously not all of these will be WANDERERS, but a significant proportion of OPAC users are clearly coming to meet short-term information needs.

At the opposite end of the scale, around 5% of OPAC users visited the collection five times or more during the last six months. These visitors have the most invested in the current OPAC, having spent time learning to negotiate it.


F&N – Have these users changed over time? (As other collections have come online etc)


The actual make up over time doesn’t seem to have changed that much, although the numbers of visitors dropped a little after a peak in early 2010.

Having said that, there are seasonal trends in the users. The search terms that UTILISERS often use to find the collections (such as “gold license”) are more popular during the school year than at other times. Similarly search terms go through peaks, depending on media interest, such as a high number of searchers who come to the OPAC looking for Australian media personality Claudia Chan Shaw, whose dress is in the collection.

Some search terms are just weird. One of the most popular search terms ever was “blue fur felt” which skyrocketed to popularity in January – July 2010, but has not been used to bring visitors to the OPAC since.

F&N – Are overseas users different from Australian ones?


During the last six months, the OPAC actually had more international users than domestic ones, with the top ten international countries visitors coming from the USA, UK, Canada, New Zealand, India, Germany, France, Netherlands and Philippines. The search terms that lead international users to the OPAC are very different from those within Australia. After all, many of the most searched for items are that link up with the school curriculum, and that is very Australia-specific. These items also make up a significant proportion of the most-looked-at references.

The search terms overseas users to access the collection are often far more specific – such as particular clock brands etc, which would indicate a higher proportion of amateur collectors (SEEKERS and FAMILIARS) than WANDERERS.

Australian users spend longer on the site, and have a far lower bounce rate, so once on site they engage more.

F&N- You’ve been speaking to our curators about how they use ours and others collection databases. What are some of the things you’ve learned from this?

Talking to the curators has been absolutely fascinating. Every single curator that I have spoken to has his or her own ways of researching and gathering collection information. Some curators rely heavily on books, while others spend a significant amount of time conducting face-to-face interviews. Others use websites like Trove, or conduct community consultation online, using wikis and blogs. However, every researcher utilises Google and the Web in some way in their search for information.

No matter how a curator conducts collection research however, all are looking for two main types of information. The first is the broad contextual information for an object that places it into an historical and social framing. This includes the broader history or biography of the creator or manufacturer, and information on the social period in which it is or was used.

The second type of information is specific to the object itself, and includes information about maker’s marks, the object’s history (including provenance, such as how, when and why it came into the collection, why it was owned and used), and any stories that relate specifically to the object.

In order to find this information however, very few of our curators use museum collection databases – even those curators who conduct a significant amount of their research online. The reasons for this varied, but emerging themes included a difficulty navigating online collections (once it could be located on the institution website in the first place), a sense of frustration at being unable to find relevant information/objects, and most important, a lack of trust in online collection databases.

Not one curator that I spoke to trusted either our own OPAC or other online collections as a resource that could provide complete and authoritative information. Where a number of curators did find online collections useful however, was in providing immediate access to images of objects and to get a sense of whether another institution held objects that might be important to their own search. Knowledge about what was in a collection was useful, but not necessarily the collection knowledge that was included in the online record.

A number of curators did use our own OPAC to see what information was being communicated to the public, and to answer public enquiries. However, it was very clear that there are ongoing issues with trust and authority.

Two things that did increase trust for curators however were good quality images (through which they could get a visual sense of the object), and PDFs of original documents. Curators trust that which they can see themselves. For most curators, their expertise is such that they will have an intuitive sense when information they come across is likely to be correct.

—

Following Susan’s initial work we started looking at the SEEKERS in more detail. Why were they coming to the site? And, more importantly, were they satisfied with what they found?

We’ve had a pop up survey running for the last two months – again using Kiss Insights – and the numbers have started coming in.

In order to survey only the SEEKERS we have set the survey to only show to visitors who’ve arrived via organic search, have visited at least three pages, and, obviously, are in the museum’s online collection. The survey, thus, has quite a limited reach and has been triggered by only 3900 visitors in the time – and has been completed by 229 respondents.

It is somewhat heartening to find that the largest subgroup of Seekers – those doing ‘amateur research, hobbyist and collectors’ – feel the content they find is ‘good’, and that the lowest positive ratings are for the ‘other’ group. This is especially interesting if we look by object and see which object records are being rated as ‘poor’. Here we find a mix of well documented (at least according to us) and very scantily documented (no image, metadata last copied from a paper stock book entry in the 1980s).

Once we get to a critical mass of respondents – 1000 or more – in this group we should have some more actionable findings. Then we move on to looking at the the other groupings.