Profile cover photo
Profile photo
Stephanie Rieger
Dormant account. Look for me at
Dormant account. Look for me at

Stephanie's posts

Post has attachment
A very insightful recap of common big data misconceptions by +Kate Crawford of Microsoft Research. 

+Bryan Rieger and I have been working this year on a healthcare project with big data aspects. We've found the following three points the hardest to deal with when planning and designing with big data in mind.

Myth 2: Big Data Is Objective
The dataset tells the story. If you're tasked with designing the product or API that will feed into this dataset, each design decision (big or small) can impact the data you will collect. Collect data from one group vs another, make one assumption vs. another, and the usefulness and relevance of the data changes. Any data may reveal insights but these will very much depend on the opportunity space presented by the data you've collected.

Myth 5: Big Data Is Anonymous
When offering publicly accessible datasets, the scale and granularity that can be visualised may reveal/infer an awful lot about personal and cultural characteristics of the participants. These may seem abstract when viewed from a distance but observers with local knowledge can often quickly discern and attribuite meaning to seemingly trivial aspects of a dataset. This can begin to reveal uncomfortably specific and identifiable details about those who have contributed the data.

Myth 6: You can opt out
Designing clear, transparent and long lasting opportunities to opt-out can be tricky, especially if data is also made available to third parties through an API. Going forward, it's also likely that almost any organisation that gathers data will find opportunities to analyse it for financial gain. The reasons to do so may not be immediately obvious (and given that no one reads EULA's regardless) opting out may not even occur to the users of a given service (nor may it occur to designers to build clear pathways to opt-out from the start).

(PS - Myth 3: Big data does not discriminate is also quite critical at the design stages. We haven't encountered this yet per se but are running into it quite a bit in our research of other products).

Post has shared content
Attn London web developers and designers with an interest in Web standards. Come and meet the W3C's TAG on the May 30th at the Mozilla office! For details "click here." #w3c

How sad is it that I just sent myself a screenshot of a G+ post because I couldn't find a way to extract or share the permalink while in the G+ iPhone app...

Post has shared content
We need more carefully deconstructed explanations such as this one. Will hopefully help focus the conversation towards achievable next steps...
Deconstructing the Internet of Things
Why the smart toaster is actually pretty cool. We tend to think of the IoT as a big sprawling beast of coordination but the early, very simple layers, have tremendous value and are much easier to get to.

Post has attachment
A few thoughts on the evolution of hover actions on touch screens...

I've been looking at demos of the Galaxy S4's Air View feature, and had a chance to try it out a few days ago.

It's hard to say whether this particular technological "enhancement" will turn out to be a blessing or a curse. Here's why...

First of all, hovering over a handheld device is hard. You have to judge the distance between your finger and the screen, which is actually quite tricky given that you are often viewing it straight-on. Doing this while stationary is quite fiddly. Doing it while in any way mobile, is near impossible. So the risk of accidentally triggering a tap instead of a hover is pretty high.

Next is the matter of intent. So far Samsung have used a hover action to "peek" at or preview data that will ultimately be shown in another view. Things like a series of thumbnails showing the images you will find in a folder, or the first x number of words in an email.

This works well, but given that Samsung can't control how people will use the feature, you can bet a fair number of designers/developers will instead use it as an alternate information layer ( not so much a "nice to have" enhancement, but to display something more critical). Bad use of a feature doesn't make a bad feature, but this certainly seems like a feature with high potential to be misused.

The last problem is discovery. Given that hoverability must be specifically implemented (and especially in the short term where most other devices don't support it)...when will you know a hover behaviour exists? This was always a problem with mouse based hovers but was a bit less critical. The effort required to hover when already holding a mouse is minimal, and there is only minimal risk of accidentally triggering an action. On a touch screen, one slip of a thumb, and you're suddenly downloading new content.

If the feature is only used to enhance, I guess it really doesn't matter whether you discover it or not, but given that the content shown will still need to be downloaded, may require some CSS, and even JavaScript...if only 10% of people know it's even it really worth forcing the other 90% to download its related assets?

Here's a demo (towards the end, the Samsung rep shows the settings screen that you can use to customise hoverability...not sure most people will realise they can even do that...)

Samsung Galaxy S 4 Air View & Air Gesture demo

Post has attachment
This almost-human-looking display mannequin is really quite lifelike. Interesting as well that to make her seem more human, she's sitting quietly  holding a purse and spends time glancing down at her mobile. 

How often do you see someone standing by themselves in a shop these days, and not passing time by staring at their mobile? That one behaviour probably causes many people to walk right by her (which in this case isn't such a bad thing...the goal is ideally not to say "hey look at our creepy robot" but "hey, that lady's wearing a nice dress")

Post has shared content
Priceless! (...and what ever happened to "commands delivered in business English". I guess that's what Siri is for?)
I always attempt to divide by zero.
Univac (Remington Rand Univac), 1956

Saw a man at Starbucks unpack a netbook, iPad and Galaxy S2, line them up nicely on the table, and settle in for some work.

Too bad you can't yet drag/beam stuff from one to another. I guess that's what cloud services are for but it still feels like we're missing an opportunity for ad hoc, proximal transfer of random stuff from one device to another.

Even cloud services like Dropbox feel inspired/constrained by the business roots of computing. "move file x to folder y".

In real life, you just randomly move stuff to temporary, spontaneously defined locations that ate only constrained by physics.

Wonder when computing will achieve that level of spontaneity?

So here's a nice example of a "feature" that in the space of just a few months has become indispensable to the point where I now judge all similar services based on the presence (or absence) of this feature and get quite impatient if the level of interaction isn't what I expected.

It also makes excellent, well balanced use of tech...the most appropriate tech for the most appropriate audience an interaction.

This tiny feature has already greatly improved my life (mostly through gains in productivity) and I suspect it's also already paying for itself on the side of the company that took the time to implement it.

It's also a piece of functionality that is just right for the audience. It uses tech that almost everyone has (SMS, email, web) and although i'm sure an app may be in the works, they also wisely understood that most users will be occasional, and therefore won't be bothered to download their app.

When and if the company does decide to create a native app, it can be ideally suited to the small subset of their audience that actually use the service frequently enough to benefit from "pro" features. A perfect native app use case.

So what's the feature? SMS and email updates providing me with a 1 hour bracket of time in which a package will be delivered. And as I found out today, the updates are in real time so if the package is unavoidably delayed, they send a new text. You can also text back to spontaneously reschedule the package to another day.

I can now plan my day, and they can now be pretty sure that once they finally show up, I will actually be home to receive the package. 

And all this with 0% effort on my part. No registration, no sharing of personal info, no download of an app. No pain to unsubscribe.

(The service in question is DPD...a relatively small UK firm. It's size could be a factor in why it's been able to implement this when Fedex and other large courriers still haven' least, not in my area).

Post has shared content
Better late than never...a link to Bryan's latest musing/sketch/hack on content adaptation for mobile and other varied contexts. As usual, it's a work in progress. It's also fitting that Bryan tweeted this a few days ago:

"It takes about a year to let an idea reach an obsessional state so I know what to do with it." - Stanley Kubrick”
Bento: - a server-side (PHP), content adaptation hack that enables designers and developers to adapt web content based on actual device features. 
Wait while more posts are being loaded