Nature Locator proposes to support the creation of crowd-sourced, geo-tagged, image-related data to enhance the ease with which quality data may be obtained for analysis by researchers. The project is funded by JISC and is a collaboration between the Web Futures group of the Institute for Learning and Research Technology, researchers at the Universities of Bristol and Hull, and members of the general public who are interested in the Conker Tree Science project.
- A new and improved mechanism for deriving automatically geo-tagged photographic evidence of the spread of this horse chestnut leaf disease.
- Sustainable integration of crowd-sourced data with analysis tools used by scientists.
- Easy-to-use, Web-based, data visualisation software to facilitate crowd-sourced
verification of the quality of the original image data submitted.
Crowd-sourcing geo data
The horse chestnut leaf miner is a tiny moth that reaches huge infestations and causes major leaf damage to horse chestnut trees in the south and east of the UK. It arrived in the UK in 2002, and has spread rapidly. The horse chestnut leaf miner is not a proven threat to the persistence of the horse chestnut trees, but may make the trees more susceptible to the lethal bleeding canker. Scientists are interested not only in the spread of the disease but also in discovering whether a natural predator of the moth –parasitic wasps – is capable of controlling levels of infestation.
During the summer of 2010, scientists Dr. Michael Pocock (Bristol) and Dr. Darren Evans (Hull), asked the general public to:
- Observe evidence of the leaf miner infestation and take part in the National survey of its spread by completing an online form with the details. The online form supports capture of geo data (latitude and longitude) by allowing the Web user to pinpoint the tree they have studied via use of Google Earth.
- Conduct an experiment to help assess the natural pest controllers (parasitic wasps) of the leaf mining moth. This was the first time that this had been done for Britain. People were asked to pick a leaf from the tree, place it in a plastic wallet, and identify the insects emerging from the leaf after two weeks (to determine how many leaf mining moths, and how many natural pest controllers were present), and similarly to enter their location and their results online.
Over 2700 people submitted results to the studies and the results presented a nationwide snapshot of the current state of the moth and its parasitic wasps This is genuine scientific research that would have been impossible without the public’s involvement.
The survey was publicised nationally in several ways (for example, a feature on the BBC’s Autumn Watch Unsprung in October 2010), but among the survey data submitted Michael Pocock and the team discovered some improbable results; they quickly realised that photographic evidence would be needed to help verify the accuracy of publicly submitted evidence of leaf damage.
How Will Nature Locator Help?
We will develop a mobile app to enable members of the public to easily identify leaf infestation, and to upload a photograph from their mobile device to a Flickr group (or similar) that will be created for the purposes of this study. This will enable auto-collection of geo data (that end-users may further verify using Google Earth if they wish) as well as time-data (allowing for more complex analysis of the data) and easy tagging of the data (for example to allow users to specify a level of severity of leaf infestation). We will also maintain a desktop, Web-based option for traditional digital camera users (i.e. those not using mobile devices). Members of the public will also be able to submit data about the parasitic wasp.
Currently Michael Pocock receives emailed photographs from members of the public, but this scenario is not sustainable: it requires manual opening and saving of each picture, only allows one person to confirm the accuracy of the photo and it also means the metadata he receives is in an unstructured format (i.e. descriptions of the location and severity of the leaf infestation are simply recorded in
the text of the email). We will provide an easy mechanism to upload the photo metadata to data analysis tools – namely the “R” freeware software package for statistical computing, commonly used by researchers in Biological sciences at Bristol.
We will prototype software that offers an intuitive and detailed visualisation of the data collected by the public on a map of the UK, to help lay people to understand more easily the spread of the disease, the parasitic wasp predator, and, importantly, to contribute further crowd-sourced data in verifying or correcting others’ identifications of the leaf infestation and levels of severity. We anticipate this will offer a kind of ratings system to allow people to vote up or down others’ estimations of the severity of leaf damage based on image observation, or even to correct erroneous claims of miner moth disease.