And so the end is near…

#Sinatrareference. Sorry about that.

Anyway, this is my last day (sad times) and because of some complicated mathematics involving cake, tea and blog posts, I have been encouraged to write down everything I can possibly think about regarding my internship. So here goes…

I spent my summer learning just about everything about Zooniverse, which in case anyone still isn’t aware is a crowd sourcing site for science. I started reading and classifying tweets, before moving onto forums and then somehow stumbling into writing a paper about the design and evolution of Zooniverse itself.

It was pretty scary at times, not in the least because being presented with a huge list of tweets and/or assorted forum posts is quite intimidating and there’s only so many “why haven’t I seen a Honey Badger,” related posts one can read through before insanity sets in. But it has been an amazing opportunity to do something ‘new’ while using my pre-existing skills and developing new ones. In particular, being presented with so much stuff to analyse, classify and summarise is very much a sink or swim moment. I appear not to have drowned, so I guess we can say I swam.

Paper writing was particularly exciting. I’d never written a paper before, having only written various essays. And while learning how to write a paper, I also had to learn how to use Github and LaTeX, the former of which was incredibly simple and the latter of which confused me a lot, because “…” somehow works out as “…”

Paper writing was intense and at times I wanted to hide in a corner and scream at things, but having a nice shiny paper (and a LOT of sleep) at the end of it was totally worth it. And now I can say I’ve written a paper, which is pretty exciting if you’re sad like me.

I’ve really enjoyed interning and while I’m excited about embarking on a PhD, I will be sad to leave everyone. And my bay. And my duck. In fact, I might steal my duck. It’s a perfect plan. Nobody will suspect a thing! Until they read this blog.

But I’ll still be around and apparently I’m an expert on Zooniverse now, which is an excellent excuse to come back and hinder/help/both with more research.

Until then though…Bye for now!

Neal

implementation!

Hi! It’s Yuki again!

My apologies for my delayed blog entry. As I have mentioned previously, I am currently developing, coding, testing and evaluating a new crime report app. Here are some screenshots of the app.

When a user opens the application, the map shows the incidents updated incident from all the users. He/she can log a new incident easily by clicking the “Incident Report” button. The user is allowed to enter a number a of things including a picture, the type of incident, the location and the time it occurred. In addition to this, the user is able to report  the incident in more detail via using text input. The incident will be saved at the our server anonymously, so that people are encouraged to post more incidents by protecting their privacy. I am still in the development with the aim to implement more advanced systems such as the filtering of spam posts , while at the same time I would like to evaluate the application. Hopefully I will soon be able to publish it in Google Play!

 

We are still here . . .

So it has been brought to our attention that we have neglected this blog a little but I promise it is because we are working hard. Anyway a quick update on what I have been doing . . .

I have been reading a lot around online forums for a paper about Zooniverse. I have been working on my Mechanical Turk study. I have researched elements of Gamification and Motivation.

This is a brilliant video about Motivation by Daniel Pink:

I would highly recommend this book for a clear explanation of Gamification and game features:

http://www.amazon.com/Reality-Broken-Games-Better-Change/dp/145583291X

As usual there are different approaches, definitions and explanations of gamification elements and features. Deterding et al explain that some researchers intentionally do not define gamification as it is an evolving area and a constraining definition could cause ambiguity and vagueness as the scope changes.

Click to access MindTrek_Gamification_PrinterReady_110806_SDE_accepted_LEN_changes_1.pdf

I have created several mock ups for this new Mechanical Turk interface and I’m currently in the process of building these pages.

I really like this tool which allows you to build attractive interface designs quickly:

https://moqups.com/

So I had better get back to fIghting with CSS and my screen layout. We will try and update you soon.

Let’s visualise data

There are many services today that allow you to track your activities, like Foursquare, MyFitnessPal, RunKeeper, Moves, Nike+ Running, iBiker, Momento, Digifit, iTrackMyTime, Mymee, Sleep Cycle… And devices like Fitbit, Nike Fuel band, Jawbone Up, Amiigo, Basis, BodyMedia FIT, Looxcie, Memoto, Muse, Shine, WakeMate, Zeo… In total, there are more than 200 devices and services you can use to track your activities. The aim is to help people to keep better track of their lives, presenting the data they collected and allowing them to set some goals, assisting them to change their behaviour, for example.

Millions of people use these services, but they are not really effective. For example, if you forget to wear your Fitbit in the morning you went to run, it will mess with your statistics, and you can’t correct it! If your Fitbit stopped working, and you got a Nike Fuel band, all the data you had from your Fitbit will not be used anymore, and as the data visualisation is different, it will not be easy to compare your activities, or just browse your activities history during the last few months.

What I am doing is an application that allows you to explore, annotate and correct the data you collected about your activities, in a personal data store (so all the data you have is stored on your computer, it’s private), and it doesn’t matter from where the data came from. Cool, right?

I’m using D3 to build some of the visualisations. D3 is a javascript library to manipulate documents using data. It can sound boring, but it allows you to create amazing visualisations! Look at that:

Soon I will post more reports of my progress! This project is really interesting 🙂

See you!

Gustavo

Language Learning

Hello, it’s Neal again.

I graduated last week with a BA in Language Learning, hence the terrible pun in the title, as I have been learning another language! MySQL.

It’s been an interesting experience, as other than messing about with C++ and Visual Basic at various points in the past, I have no experience of programming languages. I was kind of expecting some kind of nonsensical list of commands which were virtually impossible to comprehend unless the user is part (or all) robot, but in fact, I was totally wrong. It all seems fairly self-explanatory, with the exception of the “IN” command which doesn’t mean ‘in’ in the sense I’d expect it to mean ‘in’. But I’ll live.

While waiting to get my hands on a database in order to apply my new-found knowledge, I’ve been writing up the classification I made of the 17 Zooniverse projects for Elena. At first I got a bit confused – How do you write up a table without just repeating what’s in it? But then I realised that it’s what the table implies but doesn’t say that is actually the most interesting part. I’ve never really written a report before (since essays aren’t really the same) so it was an interesting experience, but I managed to write two pages worth of things which I hope everyone else will find as interesting as I did. One of the more interesting observations I made is the use of images as an asset for even audio-based projects, to encourage users to complete tasks and simplify the task for them. Spectrogram images make the task much simpler, as users can spot patterns.

Oh, and there’s a German Galaxy Zoo translation! Exciting, no?

Galaxy Zoo…Auf Deutsch!

How do you improve your interface of apps?

Konnichiwa (“Hi” in Japanese ) this is Yuki. I had my graduation ceremony on Monday, so I finally officially hold a BSc degree in Computer Science – feeling amazing! I wore Japanese traditional academic clothes called “Hakama” and top the British academic robe. It was a very hot day but I enjoyed a lot. 🙂

Currently, I am working on designing the interface of a crime report app, with which anyone can easily report a crime from their smartphone/tablet. Thanks to active discussions, I now have a number of ideas regarding the interface, such as text-based, tree-based and tagging-based. I didn’t think there were so many varieties of options! BTW, I used an online wireframe tool “Cacoo”  to make wireframes. I really like this online-wireframe tool because it allows me to easily show my ideas and share them with people! Here are some of my wireframes.

They look different, however the input factors are the same. Now I need to consider about not only quantity of the crime report but also quality. This app should encourage people to report useful information but not spam. So now, we need to think how to get useful crime reports. This is quite an interesting topic and I guess I need to research many kinds of aspects, such as HCI and psychology. It is just only the beginning, and I look forward to seek this topic and I hope I can (b)log the progress next time!

Have I bitten off more than I can chew?

Sorry for the slight delay posting, this week has been a little crazy with graduation. So my latest news . . .

So to further my role as the usability/design guru I have been allocated a new project. I am exploring what effect the user interface has on Turkers (paid participants on Amazon’s Mechanical Turk project) in relation to accuracy and efficiency of results. This sounds like a fantastic opportunity to further investigate the impact of design. The idea is to “gamify” the interface to add extra motivation incentives.

I have spent the last two days reading papers about crowdsourcing and the recent papers on mechanical turk. Even in this early stage of research I am surprised by several of the paper’s finding, including the fact that the combined efforts of non-experts can be just as good as experts in accurately completing tasks (http://www.aclweb.org/anthology-new/W/W10/W10-0701.pdf). I also hadn’t considered the different ways that users can be exploited (possibly a strong word) to complete tasks which computers currently cannot do. Several of these tasks invovled pairing up with a random stranger online and trying to guess the same labels for a photo or guess the word that the other is trying to describe. For a full explanation on these gamification models see Designing with a Purpose (http://dx.doi.org/10.1145/1378704.1378719). So before I go on for pages, repeating all my reading I should probably say what else I have been getting up to.

I have signed on to CodeSchool to try and familiarise myself with angular.js, more jQuery, backbone.js (just to name a few). I would highly recommend it (apart from the fact that you have to pay). However, for a quick month of intensive learning it seems pretty worth it (http://www.codeschool.com/). I would also recommend having a look at Codecademy which is free and I think a fantastic tool. Unfortunately it didn’t have the range of courses that I required but for the well known languages, definitely worth a look (http://www.codecademy.com/).

I have also started . . .

It is really easy to read but I have to admit I’m not as far as I would like as I have been reading all this other material as well :p.

There is more to say but even more to do so I should probably get back to it. In my next post I will let you know how I am finding Mechanical Turk and whether the implementation of this user interface is quite as traumatic as I am starting to fear it might be.

P.S. If anyone has any useful tips about implementing dynamic pages on Mechanical Turk feel free to comment below =)

Coding, characterising and completing…

Hi, it’s Neal again.

I’ve completed the rather epic task of coding and cataloguing the tweets which I mentioned last time – One spreadsheet alone contained 949 tweets. Ouch!

But it was a very useful task! Not only have I gained a feel for who the more important contributors are (their handles, but also their identities – Who they are as people), I’ve also gained a better understanding of what they’re tweeting about. Without giving too much away, a lot of it wasn’t terribly exciting but a breakthrough came today when I identified a set of tweets which formed a discussion which appears to have resulted in contributions to the discussion forums – Contributions which would assist other users with their activities both with the objects discussed and with other objects. This is a lot more exciting than it probably sounds on paper and I can’t wait to turn my attention to the discussion forums to see the relationship between the two resources.

Additionally, I’ve been characterising and analysing the 17 (!) active Zooniverse projects. This involved reading a paper and using that paper as a basis for drawing out information about different aspects of the projects to separate them from each other. Given how similar many of the projects are, this has been easier said than done and has required a bit of creativity. Fortunately, I’ve been doing this with Anna, so I haven’t had to do it alone. It’s been a great opportunity to get to know the various projects and it gave me a rather exciting excuse to transcribe some ancient Egyptian papyri, which were written in ancient Greek. Not that I’m blowing my own trumpet or anything.

I have a meeting on Monday to decide what we’re going to be doing with the Zooniverse data, so what this space:

(Don’t watch it too closely though. There are 3 more blogs to go yet!)

And seriously, look how pretty some of these research objects are!

Touching

Hi again!
I have been working for a week on an application to manipulate temporal data using touches. It’s a web application, so I’m using web technologies (mostly JavaScript), and fighting with mobile browsers. The objective of this application is to allow people to handle personal temporal data in an intuitive way, just interacting using gestures and touches.
I tried to use some well-known libs to handle the gestures, like Hammer.js, but I found that these libs can be limiting, because usually they are used just to give some touch functionalities to normal websites. All that I wanted was to know the position of each touch on the screen, and to be notified when the user releases any of the touches. I couldn’t get it using these libs (and spent a lot of time trying to do it), so I had to implement it by myself, and now it works 🙂 . I had a lot of trouble with the mobile browsers, and decided to drop the support to Google Chrome (mobile). Every time you do some touch event, it flips a coin and decide if it is going to fire the event for you or not… I hope they solve it soon. Interestingly, it works flawlessly on Firefox. I didn’t have any problem with mobile Safari as well.
Now I’m working on the real part, that is rendering the graphs and manipulating it. Let’s see what I can do 😀

You can try it on my server: http://www.gbuzogany.com/temporal/ 

Update: Problems with Firefox! Reported a bug 🙂 Giving more chances to Chrome and trying to model the coin-flip algorithm that it uses to handle touches.

Update 2: Now some problems with Safari! It’s losing some events… ahhh, browsers.

Update 3: Chrome bugs related to touch events . Things getting ugly.

Gustavo

Categorise?

Hi this is Yuki again! We have survived the first week and I am getting excited working the 2nd week. As I mentioned I am working to create new social machine crime report apps. The app should be easy to use and convenient for end users. I’ve come up with the idea for the interface to be based on decision trees. So now I am researching how to categorise. There are many type of crimes, such as Burglary, Robbery, Drugs etc….

I used WYSIWYG and an online mind map app. It is very easy to use and 3 maps are available for a basic user.
http://www.mindmeister.com/
Hoping I will sort out the categories and create WYSIWYG SM apps soon!