Crowdsourcing Objects repurposing the 1980s arcade console for scholarly image classification

Adam Crymble

Submitted Entry for 2015 Competition

Abstract

I want five minutes of your time for a crowdsourcing project. Specifically I want the five minutes you spend staring into space in the lobby of the British Library, as you eat your sandwich, or wait for your friend to arrive. The great promise of crowdsourcing is that it captures a little effort from lots of people. To facilitate this, crowdsourcing websites are accessible from anywhere with a wi-fi connection. And we all have moments of downtime that we could devote to the greater good, classifying images or transcribing text. But we don't. Everything we need seems to be there, but instead we sit and slowly chew our sandwich, eyes unfocused, mind blank.

What if the very fact that websites are always available is the problem? What if, because you can volunteer whenever you want, you feel no impetus to do so? Or maybe you're spoiled for choice. The crowdsourcing website is only one of millions of options online. You could log into your Zooniverse account and classify Victorian-era images on 'Science Gossip', or you could watch that video of a cat playing with a cardboard box.

Inspired by the 'maker' community and physical computing, my project takes the crowdsourcing experience off the web and puts it into a 1980s-style arcade game, replete with joysticks and plastic shiny buttons. If you want to play, you have to go to the machine. But unlike the Internet, the machine is put where you're waiting around - ideally as a temporary exhibition in the British Library lobby. The lobby is already home to a touch-screen interface where visitors can explore high-resolution scans of the 'Book of Golf' and other medieval works. This project takes the 'gamification' of many crowdsourcing projects, such as James Baker's 'Picaguess', and puts it into the physical game console that will trigger nostalgia for many a passer-by. The game console also encourages people to touch something physical in a building that, in an effort to protect the collection, tends to discourage tactile encounters.

This project will focus on a subset of the Flickr collection at the British Library, asking users to make some simple classifications as they browse through the images. However, the primary goal is to experiment on the volunteers themselves. If we put an arcade game devoted to crowdsourcing in the lobby of the British Library, would you focus your eyes, walk over, and play?

Assessment Criteria

The research question / problem you are trying to answer

Please focus on the clarity and quality of the research question / problem posed:

Firstly, the project provides what I believe is the world's first experiment on people's responses to 'crowdsourcing objects', which are limited in space and time, and which can only be used by one person concurrently. This will allow me to test assumptions about the value of putting material online to be transcribed or classified. Is the Internet a liability for crowdsourcing, and are 'crowdsourcing objects' a better way? The project draws data curation off the web and into the physical world in ways that are as yet uncommon and provides an excellent opportunity for experimenting with digital history interfaces, drawing on the experience of the 'maker' community and escaping from the current 'flatland' norm of the web. It acknowledges that people increasingly associate their computers with work, and by providing a digital experience that doesn't feel like a computer, we can tap into energy currently reserved for play.

Secondly, by asking users to help classify a subset of the collection in ways that are easy for humans but difficult for computers, scholars will provide useful datasets for future research on visual culture. The results of the crowdsourcing also provide new data inputs for future studies on whether or not a training set of 'caricature and outlandish' or 'carefully drawn and lifelike' images can be used to classify the rest of the collection, which could be the basis of a future grant application.

Possible classification categories might include:

  • Rate the realism of the image, from caricature and outlandish, to carefully drawn and lifelike.

  • Highlight the most interesting point in the image

  • Tag: animal, vegetable, mineral, machine, etc.

Please explain the ways your idea will showcase British Library digital collections

Please ensure you include details of British Library digital collections you are showcasing (you may use several collections if you wish), a sample can be found at http://labs.bl.uk/Digital+Collections

The project is built to showcase a sample of the British Library Flickr collection, and to do so within the walls of the British library itself. Unlike the King's Library, which is visually prominent for visitors, the digital collections are more difficult to foreground. While this project proposes to use images, once the interface is built it would be possible to use other collections as well. As none of the images in the Flickr gallery are subject to copyright restrictions, they are an ideal candidate for this project.

Please detail the approach(es) / method(s) you are going to use to implement your idea, detailing clearly the research methods / techniques / processes involved

Indicate and describe any research methods / processes / techniques and approaches you are going to use, e.g. text mining, visualisations, statistical analysis etc.

The project development will involve four stages:

Firstly, I will curate a random and manageable set of images from the Flickr collection and download them using the Flickr API. As it is the intention of the project to feed back the crowdsourced data into the metadata available to the Library, I will need to store these downloaded images along with their unique identifier in a database on the machine to be used for the arcade game.

Secondly, with the help of Ben O'Steen of the British Library Labs team, I will build an HTML-based interface containing the crowdsourcing game that will run on a locally hosted server such as MAMP (https://www.mamp.info/en/ ) or LAMP. This means that the game will not need to be connected to the Internet, dramatically decreasing load time of new images. This interface will be built to use native HTML form technology for writing user responses to the database, and will use HTML links to navigate between pages. JavaScript or PhP can be used for integrating any logic into the site. CSS will be used to make the game look professional and attractive.

This 'game' would effectively be a locally hosted website that is connected to a database, which will allow the project administrators to define some data collection fields, and will save the users' responses. This in itself is a valuable tool which could be downloaded by any scholar working on structuring information in a set of data. For example, someone looking to identify the weather in a criminal trial account could download this tool, substitute their criminal accounts for the Flickr images, alter the names of the fields of data to collect, and use the tool while they sat in an airport lounge or on a train, to classify their own documents / images. To facilitate this, the code will be released on GitHub along with a tutorial to encourage reuse.

Thirdly, the input controls in the form of buttons and a joystick would be integrated with the computer and programmed to perform the desired fuctions using JavaScript. I understand Ben O'Steen has experience with this type of work. I have contacts in the maker community who could act as paid consultants as needed.

Finally, the project needs to be built into a housing that looks like an arcade game, and decorated accordingly, before being installed.

Please provide evidence of how you / your team have the skills, knowledge and expertise to successfully carry out the project by working with the Labs team

E.g. work you may have done, publications, a list with dates and links (if you have them)

I have worked on a number of scholarly web development projects, using a range of technologies, including HTML, XML, CSS, JavaScript, and I am familiar with a number of content management systems, including Drupal, Wordpress and Omeka. I have designed a number of websites, and I am familiar with wireframing and integrating CSS design. I both designed and implemented my personal research website (http://adamcrymble.org/ ) in 2013. I have extensive programming experience for historical research applications, particularly in Python.

I was responsible for building the back-end of 'Privy Council Papers Online' (http://www.privycouncilpapers.org/ ), which involved some database development. I was also the project manager for British History Online during its recent major redevelopment (http://www.british-history.ac.uk/ ), which gave me experience planning and executing large multi-faceted projects, as well as working as part of a team on a technical initiative. I have used local development environments in the past, but my knowledge of server stacks is limited, and my abilities integrating forms with databases is also limited.

I have some experience with physical computing from my work as a research assistant in William J Turkel's 'Lab for Humanistic Fabrication' at Western University, Canada (2008-2009). In that role I learned to programme 'Phidgets' (http://www.phidgets.com/ ), which are USB sensors and buttons that either sense something (movement, temperature, light, etc), or allow physical interaction (buttons, switches, etc).

Please provide evidence of how you think your idea is achievable on a technical, curatorial and legal basis

Indicate the technical, curatorial and legal aspects of the idea (you may want to check with Labs team before submitting your idea first).

Technical
As outlined above, I believe this project could be achieved using simple website technologies (HTML, CSS, JavaScript) and a local development environment (MAMP or WAMP). The buttons and joystick are cheap and can be purchased online for a few pounds (kits available for approximately £15). The computer used to run the arcade game could be recycled from an old machine being taken out of service, as the technical needs of the machine would be very low.

In planning the technical specifications of this project I am grateful for advice received from physical computing expert Cefn Hoile (fellow of the Software Sustainability Institute) and Dr Angela Kedgley, Faculty of Engineering, Imperial College London.


Curatorial
The Flickr collection has already been curated and made available. I would be keen to work with the curators at the library to select an appropriate subset of images for use in the project, and to help develop crowdsourced data that was useful to the needs of the library.

Legal
The images are all free of copyright restrictions. No other aspect of the project has any legal restrictions of which I am aware.

Please provide a brief plan of how you will implement your project idea by working with the Labs team


You will be given the opportunity to work on your winning project idea between June 2015 - October 2015.

June 2015
Activity described here (e.g. what, when and by who)
  • Adam to meet with the BL team responsible for the Flickr collection and curate a set of images. The questions that will be posed to the users during the crowdsourcing will then be finalised.
  • BL Labs to negotiate with the BL for a place to install the finished game.
  • Adam to speak with the BL technical team to discuss the best format for building the local development environment, which forms the basis of the game's software. Work begins on the software (collaboratively).
  • Adam to speak to a consultant about considerations needed to integrate the software with the hardware (buttons and joystick)
  • Adam to source old computer to act as the game machine itself. If the library has one, that would be ideal.

July 2015
Activity described here (e.g. what, when and by who)
  • Adam and BL technical lead continue to develop the software and to begin work with the hardware.
  • Adam to begin housing design

August 2015
Activity described here (e.g. what, when and by who)
  • Adam and BL technical lead finish software and hardware.
  • Adam to user-test machine with a focus group.
  • Adam begins designing the stickers that will cover the wooden housing.

September 2015
Activity described here (e.g. what, when and by who)
  • Stickers finalised and ordered and added to the housing.
  • Adam and BL technical lead iron out any outstanding bugs in the software or hardware.

October 2015
Activity described here (e.g. what, when and by who)
  • Game installed and data collected.
  • Adam works with the BL to promote the experiment to potential users and to the media.
  • Adam to meet with the BL labs team to reflect on the project and discuss future opportunities