Ben Scott

GBIF Globe

Mon, 25 May 2015 16:18:38 +0000

Visualisation showing the GBIF specimen occurrences on the WebGL Earth 2.0 interactive globe.

MP Report Card

Fri, 01 May 2015 00:00:00 +0000

A report evaluating an MPs performance in the 2010-2015 parliamentary session, with metrics including financial interests, expenses and number of votes.

The data comes from a variety of open data sources and websites. The financial interests had to be scraped directly from the Commons’ Register of Members’ Financial Interests, and this dataset is released as open data.

Open Data Day Hackathon

Sun, 22 Feb 2015 16:18:38 +0000

At this year’s International Open Data Day, I was part of a team hacking natural history data to tell a compelling story about our changing environment.

I teamed up with @ivvetm, @sequencefree and @alexwalters75, and after brainstorming ideas we settled on mapping protected species found in the proposed Greater London National Park.

To start, we extracted a list of protected species from the Greater London Authority’s Protected Species Report.

We then queried each of these species names against the Natural History Museum’s Data Portal and Global Biodiversity Information Facility APIs, to find occurrences within the boundaries of the proposed London National Park.

Parse the data, upload it into CartoDB, and we had an interactive map showing sightings within the park.

There were some interesting discoveries - the Peregrine Falcons at the Tate Modern were new to me, but clearly represented in the dataset.

To finish, we wanted to map the change in occurrences through time. Displaying fewer species worked better visually, so we created another map showing bat species sightings since 1900.

The maps and data aren’t perfect. The London shape geometry we used to query against needs refining, so some occurrences fall beyond the boundaries of the proposed park. But we produced a really satisfying result for a such short hackathon - and had great fun doing so. Thanks to everyone involved, and especially Team Nature!

Crowdsourcing the collection

Sun, 18 Jan 2015 16:18:38 +0000

Once a year the Natural History Museum hosts Science Uncovered, an evening celebrating its science and research with visitors. My team - Biodiversity Informatics - has always taken part but since most of our work is computer-based, we often struggle to find ways of engaging the public with a monitor screen. That becomes even more of a challenge when we’re competing with an entire Museum of exciting science to explore! This year I wanted to try something a bit different.

One of the Museum’s key challenges is digitising its vast collection and making it available online. The Museum has embarked on an ambitious programme to photograph every one of its 80 million specimens, and transcribe the information on each label (what it is and where/when it was found). I wanted to tell the story of this huge task, showing visitors the process of taking a specimen from the Museum’s collection, photographing it, copying the labels, and then publishing the information on the new Data Portal. Each step would happen in real-time and use real specimens, and the public themselves would be asked to transcribe the specimen labels - crowdsourcing performed by a live crowd.

Crowdsourcing the collection in action.

The Museum’s coleoptersists have been making great efforts digitising their specimens, and have been publishing their stunning photos on Flickr. Many of these images still required transcribing, so they made the perfect seed dataset. On the night, Museum curators would take more beetle specimens from the collection, photograph and add them to Flickr. To ensure a steady stream of photos flowing through the pipeline, I set up IFTTT to automatically push the new photographs to Flickr.

Curators preparing specimens on the stand for digitisation.

The success of the project would rely on the transcription app. It had to pull an image from Flickr, allow members of the public to easily enter information from the label, and then push the data to the Museum data portal. The app needed a simple and intuitive UI - each member of the public would probably perform just one transcription, so there could be no overhead in learning to use it. On the night we had 3 iPads running the app, each chaperoned by a member of the informatics team. We would help visitors decipher the labels, and discuss our work and why digitising the collection is important. The app was also accessible throughout the Museum on visitors’ own devices - we even produced beer mats to encourage people to transcribe a specimen over a drink!

Transcribe a specimen beer mats.

After evaluating existing free (there was no budget for this) crowdsourcing apps, I decided to build one myself. The tools avilable were far more complex than we required - with user accounts, leaderboards etc., and a correspondingly complex UI. Few worked well on handhelds (phones especially), and none could consume a live stream of flickr images, instead requiring tasks to be preloaded. Our app is a simple, stripped-back version - it’s little more than an image and multi-step form - and optimised for tablet and phones. You can try it out here - and the source code is available on Github.

Visitors using the app to transcribe specimens.

The event itself was a big success. The public really engaged with our work - so much so that people continued to use the app for days after the event - and we received some excellent feedback. The app was even demo’d to the Museum’s board of trustees as an example of digital innovation. Though the dataset created is small - we received 162 transcriptions over four hours, peaking at 3 submissions per minute - the accuracy is extremely high. In a first for Science Uncovered, visitors have made a real and useful contribution to our science. And, more importantly, learnt and had fun whilst doing so.

With thanks to: Ivvet Modinou for developing the idea, Laurence Livermore for helping organise, Libby Livermore for the photos, and members of the Museum Biodiversity Informatics, Coleopterist and Digitisation Teams.

Natural History Museum Data Portal

Mon, 15 Dec 2014 00:00:00 +0000

Lead Architect of the Natural History Museum’s Data Portal, building an open access platform to release all of the Museum’s collections and research data.

As project manager, I introduced an agile (kanban-based) approach. It was the first Museum project to be run using a fully agile methodology, and the site launched on-time, under-budget and with all functionality delivered.

We also implemented an open access approach, with the project milestones, kanban board and source code publicly available.

Science Uncovered

Tue, 30 Sep 2014 00:00:00 +0000

Transcription app built for Science Uncovered 2014. At this live event, members of the public could have a go at transcribing our specimen labels.

It’s designed to be quick and intuitive to use, so our time could be spent explaining the science (or rather the handwriting).

Richseam

Tue, 01 Jan 2013 00:00:00 +0000

I was co-founder and developer of Richseam, a music recommendation engine using real connections between artists. Artists work with those who inspire them, bands share members, musicians produce and tour together.

We mined a variety of open data sources to graph two million relationships between 170,000 musicians. The site was listed as a top-50 startup at the 2013 Pioneers festival.

Scratchpads

Sun, 01 Apr 2012 00:00:00 +0000

A drupal-based open science and virtual research platform.

Based at the Natural History Museum, scratchpads allow scientists around the world to collaborate and share their taxonomic research.