Lucy
Robinson (Citizen Science Manager, NHM) began by introducing the Darwin Centre
and the Citizen Science Programme, which supports volunteer research projects.
A Digital Collections Programme is also underway at NHM, a cross-museum
programme to transcribe labels and interpret specimens, the goal being 20
million specimens accessible online within 5 years…! So research was undertaken
on crowdsourcing, which is the enlisting of an online community to provide
direct creation and interpretation of digital data. Many crowdsourcing
platforms exist and two were piloted, Herbaria@home - for a ‘Take notes from
Nature’ project focusing on accession register transcription, and Zooniverse -
which was chosen for the Orchid Observers project.
The
opportunities of crowdsourcing were summed up as the potential to showcase
collections; the expertise of the crowd; and potential to keep visitors engaged
beyond a site visit. The challenges were the choice of platform; the need to
scale up for volume of data whilst still
providing a quality experience for crowd visitors; and the additional work
involved in integrating sourced data back to the database.
Kath Castillo (Project Officer (Scientific
Communities), NHM) went into detail about the AHRC-funded Orchid Observers
project. She explained how a 2011 journal article had explored the effects of
global warming by mapping one species of spider orchid over 100 years using
historic specimens, showing flowering now 6 days earlier. Inspired by this,
Orchid Observers aimed to map 29 species using
140 years of historic museum specimens, 45 years of field
records/photographs and new 2015 data – all created and collated via
crowdsourcing. At the same time, the communication and collaboration potential
of crowdsourcing for citizen scientists was to be assessed.
The project
was launched in April 2015 and over a 5-month flowering season, with over 1,000
registered users, they obtained: 31,253 online classifications, 2,250 uploaded
images and 1800 field records including 200 new. A talk forum also proved very
successful, with users communicating and interacting independently.
Initial data
analysis confirms earlier flowering in two species, with much still to explore
including temperature data mapping. In terms of process, crowdsourcing was key
and a great success, but intensive for staff.
The talk concluded with
confirmation that the project was also great fun for all concerned – with a
beautiful smiling bee orchid image to prove it.
In the
question session, integration of data back into the main database was confirmed
as a significant task but the Zooniverse metadata should be downloadable
directly into EMu. However, the Smithsonian’s system has been recently
redesigned to include a direct interface with EMu, and so may in fact prove a
better platform long term, with this specific feature.
Lucy and Kath
also expanded on data quality control. In this case, a mainly known pool of
users allowed for a high level of trust, but the process can be tailored, for
example running data through checkers as well as the initial transcriber, with
disagreement resulting in a flag to a curator.