Citizen Science Best Practice

Following nationally and internationally recognised protocols and standards ensures that data collected by citizen scientists can be verified, replicated, and understood, maximising its usability in the future. Determining the protocols that will be used during data collection events during the planning stage, allows sufficient time and resources to be allocated for developing and testing of tools and resources as well as training staff and participants, reducing the risk of errors. Selecting sampling protocols at an early stage can also help to ensure that data forms can be designed to capture sufficient metadata at the sampling event to comply with standards and to be able to conduct a thorough quality assurance procedure at a later stage.

Specific protocols exist for many types of collection events. In the UK, UKDMOS contains many of the UK government agencies’ protocols and standards for marine monitoring programmes carried out in the UK and Europe. Specific projects and sampling techniques can be searched for in the directory, and the resulting metadata will provide references to standard protocols used, as well as when and where individual samples are collected. For example, JNCC's Guide to Fisheries Data Collection and Biological Sampling (Doyle, A., Edwards, D., Gregory, A., Brewin, P. & Brickle, P. (2017). T2T Montserrat; a guide to fisheries data collection and biological sampling. JNCC.) is referenced in UKDMOS.  However, a wide variety of data collection protocols can be found in the directory. In instances where a protocol does not yet exists, existing standards may be used to guide the development of new protocols.  If these are developed as part of governmental monitoring programmes, they can be added to UKDMOS here, or by contacting the relevant organisation hosting the standard protocols e.g. JNCC's Marine Monitoring Method Finder.

The sampling design may have an impact on the design of data collection tools and resources needed.  For example, if a survey is conducted across a vast area e.g. different areas in the UK, using an automated data submission system would ease data submission at a later stage, rather than needing to compile paper records in one physical location. When a citizen science project intends to collect data at specific sites, it is beneficial to determine where the survey events will occur (spatial coverage) as well as when and where individual samples will be collected (spatial resolution). Research has shown that data from citizen science projects with greater spatial coverage, and an increased spatial resolution are used more widely in research as the data have a smaller size of minimum and major spatial units, facilitating their use in species distribution modelling and conservation planning.  These data with large spatial coverage may be from a series of organised events at specific locations around the country, or projects where users choose where they wish to record data e.g. anywhere in the UK.

Back to top

Data collection tools can consist of mobile applications, websites, computer programs, and physical tools such as measuring devices, collection instruments, observation tools and identification guides. Things to consider with any chosen data collection tools are:

1. How easy or difficult they are to use. The difficulty of using the tools will influence the training required to provide participants with an adequate proficiency to minimise errors during data collection. If the tools are very complex, it can also deter more casual participants from joining the citizen science project; however, offering in-depth training might alleviate these challenges whilst providing the opportunity for volunteers to upskill.
2. Cost of obtaining and distributing the tools to volunteers. Depending on the funding available, compromises may need to be made, so how this may affect the data produced should be taken into consideration. For example, if a small number of identification guides, or GPS devices are to be shared among the participants, this is likely to affect the accuracy of the records so additional on-site verification of species identifications and coordinates may be required.
3. If volunteers have to purchase their own tools. This may create a barrier to engagement for participants of lower socioeconomic status. Conversely, it may also increase repeated engagement from the same volunteers once they have the required tools.
4. Risk of damage or loss at survey sites. A contingency plan should be devised in case of sampling equipment being lost or damaged to enable the data collection to continue, for example, having a few spares of tools that are most likely to be damaged, or limiting the use of these to staff and more experienced volunteers.
5. Limitations of the tools. A common limitation when conducting marine and coastal surveys is reduced phone and internet signal. If this is an issue in the intended survey locations, a mobile application that operates off-line would be a priority. Similarly, tools need to be suitable to the environment of the survey, and people might not be willing to use their own personal electronic devices in the marine environment. For example, if collecting image data of species in a snorkel or dive survey, cameras must be waterproof to the appropriate depth and produce images of a high enough resolution to enable identification of taxa from the images.
6. Accuracy and precision of the tools. This will include determining how precise the measurements need to be e.g. recording distances to the nearest cm or to the nearest m, and how accurate they are e.g. by calibrating measurement devices to ensure the measurements collected reflect the true value.

Back to top

When organising a data collection event in which volunteers will participate, there may be restrictions associated with the sampling area. For example, if it is private land or a protected area, the relevant authorities or landowners should be contacted to reach an access agreement, and to receive prior approval for surveys and scientific monitoring. They may also provide specific rules which must be followed when accessing those areas which would need to be communicated to all participants prior to arriving on-site.  Failure to obtain the required access permissions could result in an inability to use any collected data, or could prevent the data collection event from taking place.  In addition to this, certain species require a license to be surveyed, further information on which species is available here.

Participants should grant permission for the data and media they collect to be used for research purposes, publications, reports etc and if the project wishes to take photographs to be shared, on social media for example, permission must be granted prior to the collection event. Participants may also take their own photographs or collect additional data which could be supplied to the project as supplementary data. The conditions of use should be agreed with the participants. In a permissions form, it may be useful to separate conditions for data/media access and reuse from permissions to use in social media and advertising as some participants may grant permission for data to be used in reports, btu not on social media, for example. 

In addition to this, when conducting a thorough quality assurance, the data collector may need to be contacted to confirm details of the records, so some basic contact details for participants may need to be retained until the data has been through its QA. However, personal data should be limited to only collect what is relevant and necessary. If any personal data i.e. information that relates to an identified individual, is collected a plan must be in place to comply with GDPR guidance and regulations for data sharing under the Data Protection Act 2018. This will include ensuring personal data is processed securely, potentially by using pseudonymisation and encryption to enable data confidentiality, integrity, and availability.

Back to top

Data verification methods to be conducted at the collection event and any relevant quality control (QC) schemes should be determined within the data management plan. Verification can involve the checking of records at the site by subject experts to ensure records are accurate and to reduce the risk of species misidentification, or the incorrect recording of coordinates. A trial collection event conducted by professionals carried out prior to the citizen science event can help to mitigate errors and improve accuracy by identifying potential risks, developing further training where necessary, and providing a baseline data collection standard to which the citizen science data can be compared. 

Data verification following a data collection event ensures that data and taxa identifications are accurately recorded. Verification processes can be built-in to the platform where data will be recorded, for example iNaturalist and iRecord, or it may need to be planned separately if the citizen science project intends to publish records via a different route, e.g. as a whole survey to a data archive centre. This could involve collecting samples at the site of data collection and having an expert in that taxonomic group identify the taxa at a later date.  Another example of data verification is the process the Zooniverse uses in projects where images are analysed by multiple volunteers.  This differs from data validation, which may involve automated processes to validate that data are in the correct format and meet the required data standards.

The NMBAQC scheme describes the importance of collecting a voucher/reference collection when recording biodiversity data to be used as a reference for future identification of taxa or habitats. This voucher specimen can be a preserved specimen, or clear images of the specimen or habitat. 

Another quality control method is to repeatedly sample the same area and obtain an average of results, which may be more suitable for projects where the sampling area is relatively small and there are sufficient participants to repeat sampling. Quality control differs from quality assurance (QA), as QC involved checking the data for errors e.g. bias and systematic errors, and recorded correctly prior to submission to a DAC, whereas QA assesses the data are correct, accurate, in the correct format and that it meets data standards, and may include a validation step prior to publication, and this step may be carried out by the DAC which will archive and publish the data.

By outlining the verification and QC methods which are expected to be carried out at the data collection in the data management plan, the methods can be standardised across survey sites and any additional training, equipment or staff required can be planned well in advance.

Back to top

Participants will need the appropriate skills to conduct the tasks required during the citizen science project. Outlining the training requirements in the data management plan enables sufficient time and resources to be allocated for training participants, both volunteers and staff, before starting to collect data. Training sessions may focus on the preparation for the event, what to do during the event or how process the data following the event. Training must be tailored to the complexity of the data collection protocols, equipment used and the experience level of the participants. Training may occur before, at the start of a data collection event, or after if focused on data management. The following elements should be considered when planning training resources:

1. Complexity of collection methods. Complex sampling methods will require more extensive training thus benefiting from closer supervision by either professional scientists or volunteers with previous experience with similar techniques. Complex methods are also more likely to incur errors, so reducing the complexity as much as possible reduces the risk of bias and systematic errors. This can be achieved by breaking down complex tasks into easier tasks, e.g. when recording species in biodiversity surveys, splitting the volunteers into groups and allocating a subset of species for them to record allows them to become more familiar and skilled at identifying those taxa.
2. Equipment used by volunteers. Similar to the complexity of collection methods, equipment that is more difficult to use will require further training and supervision by those who are more skilled in the use of the collection tools.
3. Experience level of citizen scientists. Volunteers that have completed more advanced training, are likely to produce more high-quality data. Less experienced participants will require more comprehensive training. A buddy or group surveyor system can be beneficial for surveying events where there is likely to be a range of abilities, pairing experienced participants with less experienced volunteers.
4. Standardisation of training resources and programmes. If the project is to be carried out in different locations, the training provided to participants should be consistent across locations to ensure the data collected is of the same quality. This can be achieved by having an overarching training curriculum followed by all those providing training to volunteers, using the same training resources wherever possible. Online training sessions are an effective way of delivering training to people in different locations, or funding an in-person workshop in a centralised location. Prior testing participant abilities can improve reliability but can be a barrier to participation if requiring significant learning time.

Training materials can consist of pre-collection training such as courses (online or in-person), assignments to get participants familiar with the collection methods and the data handling required, video tutorials, identification guides and survey technique guides. 

Back to top

Data standards are a set of rules for formatting data to make them consistent across datasets. Complying with data standards enables the data to be interoperable and reusable, so they can be shared with national and international data aggregators, used for a wide variety of work. Being interoperable and reusable means the data has a greater impact as they can be used in environmental monitoring programmes, they can inform essential biological variables (EBVs) and essential ocean variables (EOVs), and can contribute to initiatives like the UN Ocean Decade, as well as the Marine Strategy Framework Directive. 

Data standards should be determined prior to data collection as this will help to identify which data fields are mandatory to complete in order to comply with the standards e.g. survey date, survey location, sampling devices. Doing this early on means that you will be able to ensure this important information is captured as standard from the beginning of the project.

The standards selected will depend on the type of data collected and the theme it falls under. In the UK, MEDIN is the network of marine-focused data archive centres that contribute to the creation of marine data standards and implement the adherence to these. MEDIN standards are categorised by marine data theme e.g. biodiversity, oceanography, meterology, with templates and data guidelines that data providers can complete to ensure their data are compliant by completing all mandatory fields. Data that are MEDIN.compliant are also GEMINI  (national) and INSPIRE (global) compliant, which means they comply with national and international standards. These data can also then be transformed into other data standards such as DarwinCore when published to enable interoperability and the harvesting of data by aggregators such as OBIS. Other common data standards are Ecological Metadata Language and BioCASe / ABCD.

Back to top

Metadata is data that provides information about other data. It provides information to data users which help them to understand the data and how they can be used without needing to contact the data provider. When a dataset is published, a metadata record should also be published alongside it. A metadata record provides detailed information such as who owns the data, any access and reuse restrictions, any unique identifiers for the dataset, the date it was collected and published, where it was collected, where to access it and who to contact if there are any questions about the dataset. Metadata records should also comply with standards. To produce a high-quality metadata record the standards should be researched prior to collection events to ensure all mandatory information is recorded. MEDIN metadata records comply with the MEDIN Discovery Metadata Standard, and are published on the MEDIN Discovery Metadata Portal, where users can search for datasets and their metadata. 

Back to top

Once the data have been collected, they may need to be processed to put them into a format which can facilitate data standardisation, quality assurance and publication. This can involve digitising paper records, scanning documents, or identifying taxa from photographs and videos. The processing steps should be outlined in the data management plan along with any tools required. This will enable sufficient training and resources to be allocated for effective processing of the data. For example, if data will be processed using an automated script, personnel must have the appropriate skills and software to do this. Examples of data processing tools may include Microsoft Excel, Python, R package., QGIS, or coordinate conversion scripts.

Back to top

Quality assurance (QA) of data and metadata is carried out after any data processing, and aims to identify and correct any mistakes that may exist in the data produced by a project. The main areas that should be assessed during the QA of a dataset are spatial accuracy, taxonomic accuracy, methodological consistency, and temporal accuracy. Adherence to the selected data standards is also checked during the quality assurance of data and metadata, prior to validation of the dataset against specific standards to minimise corrections required at a later stage. When submitting data to a data archive centre, as much information as possible should be provided to the DAC so they can carry out a thorough QA of the data. This information can include any raw data, including original paper records if relevant, any images collected, GIS files, videos, contact details of the data provider, any relevant documentation such as project reports, vessel reports or logs. By establishing the QA procedures at an early stage in a citizen science project, a checklist can be made of information that must be retained for effective QA of the data. 

When quality assuring the data and metadata, a variety of tools could be used, such as Microsoft Excel formulae, R package., taxon match tools, e.g. WoRMS and MSBIAS, coordinate converters, geographic mapping software like QGIS or Google Earth. If submitting data to a data archive centre like DASSH, you may not need to conduct the quality assurance yourself as they may have their own quality assurance procedures. The relevant DAC should be contacted to discuss data requirements for conducting their QA procedures. A list of MEDIN data archive centres and their contact details can be found on the MEDIN website. 

Back to top

Data archive centres (DACs) archive and publish data, making them available for a wide range of users and data aggregators to harvest and use. They can also provide data-focused leadership, best practices, services, tools, and training to support the archiving and publication of data in a standardised format. This ensures data is findable, accessible, interoperable and reusable, or FAIR.

MEDIN, the Marine Environmental Data and Information Network, consists of 7 data archive centres which specialise in archiving different types of data from fisheries, water column oceanography, meteorology, bathymetry, historic environment, geology and geophysics, flora, fauna and habitats.

The relevant DAC should be contacted during the planning stage to be well informed of the data submission process, standards used, and any further requirements. If you are unsure about which DAC the data from your citizen science data should be submitted to, or have any questions, you can contact the MEDIN team at enquiries@medin.org.uk. 

If the citizen science project has received funding, it is possible that the archiving and publication of any data produced to a DAC is a requirement. Even if this is not the case, it is still strongly recommended to do this to ensure data are FAIR and their longevity is maximised well beyond the lifespan of the citizen science project. 

Back to top

The knowledge that their data will be contributing to science, public information and conservation can motivate citizen scientists to take part in a project. Determining the pathway the data will take once collected, and communicating this to potential volunteers when advertising the event could also increase the number of participants and the quantity of data that can be collected.

The type of data recorded by citizen science projects, as well as the intended degree of sharing of the data will influence the pathway those data will take. For example, single records submitted by users of iNaturalist will follow the route in Figure 2. These data are more likely to be collected by casual citizen scientists recording ad-hoc sightings, but individual record submission platforms like iNaturalist and iRecord can also be used to aid data submission from organised data collection events. Larger datasets are more likely to be archived and published by data archive centres like DASSH (marine biodiversity and habitats data); the pathway these data take from DASSH are also shown in Figure 2.

Figure 2: DASSH data flow

Case Study

The Rock Pool Project has been clear with volunteers regarding the pathway their data travels through, communicating this via training presentations prior to their citizen science events. Figure 3 shows the flow diagram they have shared with participants.

Back to top

All high-quality data produced by a citizen science project should aim to be FAIR: findable, accessible, interoperable and reusable. This can be achieved by complying with data and metadata standards, making these openly available to users in an interoperable format, and as openly accessible as possible. By researching data licenses and reuse conditions at this stage, the data pathway and collaboration with data archive centres can be started at an early stage, and to liaise with funders and data owners on how the data will be shared upon completion of the project or collection event.

If volunteers are collecting and submitting data directly (e.g. through a mobile app) then you may want to consider providing guidance/training on data licensing and how to select the appropriate license to ensure maximum usability.

Back to top

After a dataset is published or archived, data maintainers will be responsible for making amendments to the dataset itself or to the associated metadata. This may occur if the ownership of data changes, if the title changes to match a publication, if amendments are made to records, or if the further surveys are carried out in the future which are part of the same project and so can be linked together via a series. If the data are published via a data archive centre, an agreement can be made for the DAC to be responsible for the updates, or the data owners or other named organisation can carry out any updates. To ensure data remain accurate and reusable, it is important that the maintainers are specified when data are archived and published.

Back to top

To produce high-quality data, the equipment will need to be standardised and calibrated across volunteers and staff. Iterative development and design of the tools required, i.e . refining the tools required for volunteers through a few rounds of ‘test’ collection events to identify the limitations and potential systematic errors using certain tools may incur. Further information on iterative design of tasks and tools can be found here.

The data could be collected by using handheld devices such as a smartphone or laptop, they could be recorded manually on paper, or use a combination of both, where citizen scientists may initially record data manually but they submit these data online. Whether the collection method is manual, hybrid or virtual will affect which tools are required. Large-scale projects which require volunteers across a large area to submit records in a given time frame may benefit from using a hybrid or virtual collection with online tools, whereas smaller scale, supervised projects may benefit from recording taxa manually, then collecting and verifying the manual records before digitising them for collation and submission to a DAC for publication.

The tools required will also depend on the type of data the project intends to collect.   For example, whether this will consist of taxonomic classifications, percentage cover estimates, presence-absence determinations, counts, organism trait measurements or environmental measurements. A rock pooling survey where volunteers record the presence of all organisms encountered within the survey may only require the use of a recording app, or a sheet of paper along with an identification guide. However, a project that wishes to conduct transect surveys along a coastline to measure environmental and biodiversity variables like shore cover and taxon abundance might require the use of distance measuring devices, quadrats, as well as the recording tools like cameras, paper or handheld devices. In both cases, the tools should be standardised, calibrated, and assessed for suitability to the environment and skill level of volunteers.

Examples of data collection tools that may be used in a citizen science project include

• Manual recording materials e.g. notebooks, pre-designed worksheets to be handed to volunteers.
• Virtual recording materials e.g. mobile phones, apps, computer programs such as Microsoft Excel, websites such as iRecord, online forms, cameras.
• Measuring tools e.g. measuring tapes, GPS devices, quadrats, transect lines.
• Specimen collection tools e.g. nets and traps.
• Guides e.g. identification guidebooks, virtual guides, identification software e.g. like the species identification software in iNaturalist.

Back to top

The data collection methods should be appropriate for the expected experience level of the citizen scientists, with relevant training and tools provided to ensure the data is collected accurately to produce high-quality data. The methods used will depend on the survey type e.g. transect, quadrat, seine-netting, pond-dipping, rock pooling, mammal trapping, walkovers, push-netting, grab/core sampling, dive or snorkel surveys, timed species-searches, plankton netting. These collection methods might be carried out by volunteers, professionals, or a combination of both.

To ensure data collected are of a high quality, trial collections are recommended, where professional or experienced surveyors conduct the collection protocols using the selected tools to determine where systematic errors or difficulties are likely to occur. This allows for the protocols to be developed further to reduce the risk of these errors, or to modify the processing of the data to adjust for systematic bias e.g. calculating mean abundances over a few groups of volunteers. In addition to this, previous research has highlighted that with adequate training and low-taxonomic-resolution protocols, citizen scientists can generate data similar to those of professional scientists.

Standardised methods for data collection are available to ensure data collected is interoperable and reusable. An extensive list of standardised methods is available here, with a list of protocols and standards for UK and European marine monitoring available on UKDMOS. 

Back to top

Verifying data that is recorded at the collection event and conducting quality control on this and the data themselves is recommended to produce accurate and reliable data of a high quality. The amount of on-site verification and quality control (QC of data will depend on citizen scientist skill levels, design of collection procedures, equipment, expert verification, replication across volunteers, accounting for random error and systematic bias. Verification may involve having experts verify taxonomic identifications and locations prior to submitting records or checking small samples of data collected by volunteer groups to ensure key taxonomic groups have been identified correctly. Checking samples of data from individuals or groups of volunteers is also likely to reduce the effects of inter-observer variability, especially where data recorded consist of estimates rather than absolute numbers e.g. percentage algal cover in an intertidal quadrat survey.  This differs from validation, where automated procedures may be used to check the data format and its compliance with the desired data standards/

Back to top

At the data collection event, permission to use images of and data collected by volunteers should be obtained if not already received prior to the event. If data is to be collated into one survey dataset which will be submitted to a DAC, and no personal details of the volunteers are supplied in the dataset, it will be the responsibility of the owner of the data, i.e. for the whole project, to select a suitable license. However, if volunteers are submitting their own data to a website or database from a personal account, they must select the license under which they wish to share their data to enable its reuse in the future.

Permissions for accessing a specific survey site should be determined prior to data collection, in the planning stage. At the collection event, any restrictions or rules of access should be communicated clearly to all volunteers.

Back to top

Participants should have the appropriate skillset for the work they will be carrying out during a citizen science project¹. This will involve providing any relevant training resources prior to, during, and after the data collection event as well as any relevant tools, like recording apps such as iNaturalist.

Training can also be tailored to the experience level, providing more advanced training to those who are more experienced and encouraging citizen scientists to progress through the training programme and develop their skills. The experience level of surveyors can be categorised, from most to least experienced, as professional, academic, naturalist, volunteer with expert ID, or volunteer. Buddy systems can be used during the data collection to allow more experienced surveyors to check the work of newer volunteers and to develop their training further.

In general, the method of data collection can either be carried out by volunteers and classified by professionals, carried out by professionals and classified by volunteers, or both collected and classified by volunteers. Each of these methods will have different training requirements to maximise data quality. For example, if data are collected by volunteers and classified by professionals, training should focus on the use of measuring devices and recording equipment such as cameras, as professionals would conduct the analysis and classification of records. If the inverse occurs, training for volunteers would also need to focus on taxonomic identification.

¹ Kosmala, M., Wiggins., Swanson, A. and Simmons, B. (2016). Assessing data quality in citizen science. Frontiers in Ecology and the Environment, 14(10), 551-560. doi:https://doi.org/10.1002/fee.1436

Back to top

Data archive centres can aid with the standardisation and quality assurance of citizen science data. They may require datasets to be submitted in a specific format such as the MEDIN guidelines, in order to conduct quality assurance free of charge, or they may be able to provide data digitisation and standardisation services if required. The DAC responsible for archiving the relevant data type should be contacted at an early stage to discuss what services will be required from them and the data they will need you to provide in order to perform these services.

Back to top

The data verification in the processing stage is carried out following data collection, and aims to check any records for accuracy and reliability. This is particularly recommended if on-site verification is not possible, for example for citizen science projects where volunteers collect data on an ad-hoc basis rather than as part of a group in an organised event. Where mobile apps and web forms are used to submit data they may redirect records to a verification system where expert verifiers can confirm taxonomic identifications, and then have automated integrated data validation systems to ensure records contain all the mandatory information prior to submission. Data archive centres can also verify records by contacting experts in specific taxonomic groups to verify species identifications and their geographical ranges.  Verifying records prior to processing the data further reduces the risk of taxonomic and geographical errors in the data, producing accurate and reliable data.  This differs from validation where automated processes may be used to ensure data are in the correct format and comply with the specified data standards.

Case Study

Records can be submitted to iRecord directly through the website, or via iNaturalist. iNaturalist records which have had their taxonomic identification confirmed by two or more naturalists, and have all mandatory metadata, flow through to iRecord for expert level verification. iRecord has a series of automatic checks based on rules about how difficult the species are to identify, whether they have been recorded within their known distribution, and within its regular period of activity. The automatic checks help expert verifiers to identify which records may need further examination prior to submission. The records are initially marked as “Unconfirmed: Not reviewed” before experts confirm whether they are accepted, or not accepted, as shown in figure 4. The experts are usually volunteers with a specialty in specific taxonomic groups who work on behalf of the national recording schemes. This automated identification of records and the manual verification step increase data quality and ensures records are suitable for use in research, planning and conservation.

Back to top

Data standards are best practice guides for archiving marine data. They provide a framework of information that should be collected with your data to ensure that they can be reused by others in the future. Standards allow information to be captured at the time of data collection or immediately afterwards, so they can provide a checklist of essential information to collect. They instil good practice amongst users as they provide a standard format to work to and allow easy ingestion of data into data archiving centres. Some DAC. like DASSH will be able to publish data for free if it is submitted in one of the data standards such as MEDIN or Darwin Core. Data submissions in non-standard formats is possible, but there may be cost involved for transformation.

Standardising data involves converting the raw data into a format that is consistent with specific data standards, so that it is interoperable and reusable. Standardisation of data and metadata allows users to understand data effectively without needing to contact the data provider, reducing the risk of misinterpretation, saving time and resources used to reformat data into a usable form. This vital step also enables onward publication or sharing with national, regional and global data aggregators, increasing its impact by allowing more users to find and access it, supporting FAIR data principles.

Common data standards for biodiversity data include:

• MEDIN guidelines. MEDIN produce standard guidelines for a range of marine environmental data themes. Data that are compliant with MEDIN standards, are also compliant with UK GEMINI, EU INSPIRE Directive and ISO 19115 and 19139 data standards.  
• Darwin Core. Data submitted to NBN Atlas, OBIS and EMODnet must be provided in Darwin Core format, which is an international standard for biodiversity data.  
• ABCD is the Access to Biological Collections Data exchange schema, a biodiversity data standard for the access to and exchange of data about specimens and observation.
• Other marine data standards are listed on the MEDIN website.

Metadata standards include:

• MEDIN Discovery Metadata Standard. This metadata standard is also compliant with GEMINI and INSPIRE standards. Compliant metadata records are published on the MEDIN Portal.
• UK GEMINI, the Geo-spatial Metadata Interoperability Initiative, is a geographic metadata standard for describing geospatial data resources.  
• Other metadata standards are available in the RDA (Research Data Alliance) metadata standards catalog.

Data can be standardised by completing a pre-designed data template, such as the MEDIN biodiversity data guidelines, or by following specific guidance to format datasets according to a checklist of mandatory and conditional fields which must be completed for compliance. Data are recorded using controlled vocabularies, which have standardised definitions and term titles for pieces of information, removing any ambiguity within a dataset. Data archive centres can provide comprehensive guidance on data standardisation, and further information is available on the websites for each data standard. If a standard for the type of data the citizen science project will produce does not yet exists, data archive centres can also help to develop a suitable, tailored standard for your data.

Back to top

Metadata is information that describes data. As a minimum, metadata describes who collected the data, where they were collected, when they were collected, what the dataset(s) describes, how the dataset was collected, and a contact person/organization or data provider from which a copy of the data can be obtained. Metadata standards ensure metadata are collected in a consistent way which allows them to be shared or combined to form a national information resource and help to answer management questions both nationally and internationally. The importance of recording metadata is highlighted in this video. High-quality citizen science data will have a metadata instance which effectively describes the dataset making it as accessible as possible for any users.

Case Study

MEDIN metadata records are published on the MEDIN Portal. Each record gives metadata such as the abstract, data holder, use constraints, geographic extent, a link to the data location, unique identifiers, and survey dates. Users can search for specific survey data by finding their metadata record, and should be able to download the data within two clicks of finding the metadata record. Contact information is also available should the users require additional information.

Back to top

Controlled vocabularies are one of the tools used to make datasets more consistent and standardised, supporting FAIR principles by making data interoperable and reusable. They make potentially ambiguous terms explicit and clearly defined which reduces the risk of misinterpretation. For example, the common name for a species may be different in different areas, so by using a controlled vocabulary for taxon name with a standard species name and code it is clear which species the data is describing. They also allow the online publication of data in a format that is machine readable, which means that APIs and machine learning can collate data from different sources which can be used to conduct analyses.

Controlled vocabularies are used in metadata standards in fields like keywords, geographic coverage terms, formats of data files, and in data standards to describe fields such as categories and models of instrument, organisations and research vessels. Some data standards, such as the MEDIN guidelines, specify which controlled vocabularies to use in specific data fields, but data archive centres (DAC) can provide additional assistance with standardising data with controlled vocabularies.

Commonly used vocabularies in biodiversity data standards are:

• WoRMS, a catalogue of marine species that includes unique “aphiaID” codes for each taxon name
• MSBIAS, a subset of WoRMS for the British Isles and adjacent seas.
• ICES reference codes for vessel names (C17), sampling devices categories (L05), sampling devices (L22).
• Vocabularies for biological features such as size (S09), sex (S10), development stage (S11), morphology (S14), colour (S15)
• EPSG codes for coordinate reference systems
• P02 vocabulary for metadata keywords
• EUNIS codes for biotopes: M21, M20 and M24
• EDMO organisation codes e.g. for analytical laboratories and data owners

There are vocabulary search clients such as the BODC client and SeaDataNet client in which users can search across many different vocabularies in the NERC vocabulary server. The OBIS Manual has specific guidance on how to select a suitable term in controlled vocabularies, and how to map terms to specific standards such as Darwin Core (DwC), and MEDIN also provides further guidance on the use of controlled vocabularies.  Please note that this is not a comprehensive list of controlled vocabularies.

Back to top

In order to standardise data and quality assure data, tools may be required to conduct automated checks or to convert data. Projects that produce high quality data will provide any volunteers or staff that will process data with the adequate training to ensure they are competent with the tool and to reduce the risk of mistakes. For example, if transcribing data from one spreadsheet to another, there is a risk of copying errors due to Microsoft Excel’s automatic formatting. This highlights the importance of conducting a thorough quality assurance following the standardisation and processing of data into final versions. The tools themselves can also be used to check that the processing was carried out accurately, and to check for transcription or recording errors.  Examples of data processing tools, in addition to Excel formulae, include Python, R packages, QGIS, and coordinate conversion tools such as Grid Reference Finder or OBIS's batch coordinate converter.

Back to top

Following data collection and standardisation, the data should be quality assured to correct any errors that may have arisen during the recording of data or during the conversion to the standardised format. Citizen scientists may have varying levels of data and computer literacy, so errors introduced by copying values between Excel spreadsheets, converting units, and selecting unique identifiers for controlled vocabularies may occur. A thorough quality assurance procedure will increase the reliability of data, and by checking adherence to standards it will also increase the interoperability and reusability.

When submitting standardised data to a data archive centre such as DASSH, it will be quality assured by experienced Data Officers, cross checking all information, images and videos, against the original data and relevant standards, reviewed to assess data quality. Any changes made are recorded to enable traceability of modifications. If any information is missing to comply with data standards, the DAC will then contact the data provider to obtain this and clarify any ambiguities. Further information on DASSH quality assurance procedures are available here.

There are many tools that can be used to assist with the quality assurance of a dataset including:

• WoRMS (World Register of Marine Species) taxon match tool. Checks taxon names are accepted and provides the correct aphia IDs, which are unique and standardised numerical codes for each name.
• MSBIAS (Marine Species of the British Isles and Adjacent Seas) taxon match tool. The same as above but only for marine species in the British Isles.
• Tools used to validate datasets for sharing to OBIS are available in the OBIS manual.
• QGIS to check coordinates of data points are accurate according to metadata and the geographical range of species recorded.
• Grid reference finder to check conversions of coordinates from one format e.g. British National Grid, to another e.g. decimal degrees.
• Google Earth can be used to check coordinates and localities of sampling events, particularly when checking against named locations.
• NBN Atlas can be used to check current distributions of species in the UK.

Back to top

Validation of data against the selected standards ensures that the data are compliant with those standards to enable onward publication and interoperability of the data. While quality assurance checks for general data errors, validation specifically checks for standard compliance prior publication of the data.   Validation may be carried out as part of the QA step, or it can be treated as a separate step before publication.  Examples of validation procedures include:

• DASSH's MEDIN Guideline Validation tool. This tool checks that a dataset which has been standardised into the MEDIN guideline format is compliant, flagging any issues or fields which do not match the standards.
• GBIF data validator. This provides a report on the syntactical correctness and validity of the dataset, requires the dataset to be submitted in Darwin Core (DwC) format.
• MEDIN Discovery Metadata validation (Schematron). When completing a MEDIN metadata record using the online editor, the editor will run the record through a validation before it can be exported to the MEDIN Discovery Metadata Portal. The metadata records can also be exported as XML files, or can be created using Maestro, a tool which saves the records as XML files. The schematron can be used with the XML Schema Definition to validate XML metadata records.

Back to top

Data produced by a citizen science project may be digital or physical. Digital data could be spreadsheets, digital reports, databases, acoustics, digital images and GIS files. They can be supplied to a data archive centre in this format for their archival as raw files, along with standardised versions of dataset and metadata that can be published to make them accessible.  Some DACs such as DASSH may be able to archive novel data types such as eDNA data and spatial polygon data, further information can be found by contacting the relevant DAC,

Physical data such as survey sheets, record cards, photographs, notebooks, reports will need to be digitised prior to archiving if the data are to be published online. Although it is possible to archive these physical assets in repositories, museums and data archive centres as they are, by creating a digital record of them they become discoverable and usable by a much wider range of online users and organisations, supporting FAIR data principles and maximising the impact of these data.  The relevant DAC should be contacted to establish what their capacity is to store physical data. If they do not support physical data they may be able to advise on alternative archiving solutions, for example with museums or libraries. 

The archiving of these data with a data archive centre is often a stipulation of funding grants, so it is highly recommended that this step is carried out promptly after the creation and processing of the data. It also ensures the longevity of the data, especially in cases where projects conclude, where organisation that own the data dissolve, or where individuals who own the data are no longer around.

Back to top

When citizen science data are published and made openly available for re-use, they will have a greater impact as they’re more widely available for use in research and data-driven policy. This impact is maximised when data are FAIR because they’re published in a findable location like a data portal, under an open-access license, are standardised and well described with standardised metadata, and are in a format that can be reused and understood clearly. Similar to archiving, the publishing of data produced by a project funded by grants is often a mandatory requirement at the end of the project.

Back to top

Data archive centres fulfil the vital role of preserving data by archiving and publishing it. High quality data collected for UK marine assessments should be archived at a nationally or internationally accredited data centre to ensure long-term access to these data, e.g. MEDIN Data Archive Centre, ICES, other data centre accredited with Core Trust Seal.

Back to top

After the archival of data, it must be shared in order for it to be found and used for research and data-driven policy. Publishing can be done through accredited data archive centres, and should provide, where possible, general metadata which describes the dataset, and a way to download the dataset itself. Published data can made discoverable on data portals, via published articles, and via downloadable or searchable databases. In each case, access and reuse conditions, stated in the form of a data license or named conditions, must be specified clearly to enable open data by third party users⁴.

Data can also be published by submitting individual species records to data portals via apps and websites. This method is commonly used by citizen science projects where volunteers submit sightings on an ad-hoc basis or as part of a surveying event organised through platforms like iNaturalist. In this instance, data licensing and reuse conditions must also be specified for each record to allow collation and analysis of these data in the future.

⁴ Bowser, A., Cooper, C., de Sherbinin, A., Wiggins, A., Brenton, P., Chuang, T.-R., Faustman, E., Haklay, M. (Muki) . and Meloche, M., 2020. Still in Need of Norms: The State of the Data in Citizen Science. Citizen Science: Theory and Practice, 5(1), p.18.DOI: https://doi.org/10.5334/cstp.303

Back to top

Data archive centres are able to publish and share data on many platforms, with an extensive network of data users, provided they are in a standardised format. They provide datasets with a DOI, a digital object identifier unique to the data, and make this available on data portals, global databases and metadata portals. A DOI allows traceability of data, it is referenced when used, and can be versioned to reflect changes or updates in the dataset. DOI. have a landing page giving basic metadata about dataset, and the link to download the dataset itself. Either the DAC. or the data providers may create the MEDIN metadata record that references the specific DOI for that dataset. DAC. like DASSH may quality assure a metadata record created by a data provider to ensure it meets the metadata standards, and adding the DOI information once created, before publishing the record and making the data discoverable. When provided with permission to do so, DASSH also promotes the publication of new datasets by sharing news of these publications via their X, formerly Twitter, account, @DASSH. The relevant data archive centre can provide specific information about the publication process, the list of MEDIN data archive centres and their contact details can be found here.

DACs can make standardised data available on an Integrated Publishing Tool (IPT), which is an open-source software used to create and manage distributed data repositories that share data across a network. National, e.g. NBN Atlas, and global, e.g. GBIF and OBIS, data aggregators can harvest these data and make them accessible via their data platforms, vastly increasing the number people who can find and access these data.

Back to top

When sharing data produced from a citizen science project, the route through which the data subsequently flow will affect how widely accessible they are. This route is determined by factors such as whether or not the data are in a standardised format that can be harvested by data aggregators, and whether they have metadata records describing the datasets so they can be interpreted correctly and so provenance of any data used in research can be tracked and credited. MEDIN DAC. have established data flow pathways, so by publishing through a DAC data providers can ensure their data is openly accessible to a wide range of users.

Back to top

When publishing data, a license will need to be specified to inform any data users of the reuse conditions for the dataset. It is also possible to list specific reuse conditions which outline what purposes your data can be used for and by who. Making data originating from citizen science available under open access licenses makes it reusable and maximises the impact the data can have, such as informing policies^6.. Data published without a license or explicit terms of use cannot be considered open data, so it is vital to specify these to produce high-quality, FAIR data. Common license types are:

• Open Government License (OGL). Used to enable public sector bodies to license use of their data and use the data, subject to one single condition. This license allows you to:
  • Copy, publish, distribute and transmit the data
  • Adapt the information
  • Exploit the data commercially and non-commercially
    However, you must:
  • Acknowledge the data source and include any attribution statement specified by the Data Provider(s) and, where possible, provide a link to the license information.
  • More information on this license can be found on the National Archives website.
• Creative Commons No rights reserved license (CC0). Under this license, the data provider has waived any rights to the data under copyright or Intellectual Property Rights law. This license allows you to;
  • Copy, modify and distribute the data, including for commercial purposes, without seeking permission from the data provider
  • More information on CC0 can be found on the Creative Commons website
• Creative Commons license with attribution (CC-BY). Data available under this license can be used for any purpose, including commercial use. This license allows you to:
  • Share - copy and redistribute the data in any medium or format
  • Adapt - remix, transform and build upon the data
    You must:
  • Give appropriate credit to the data provider
  • Provide a link to the license
  • Indicate what, if any, changes have been made
  • More information on this license can be found on the Creative Commons website
• Creative Commons with attribution non-commercial (CC-BY-NC). Data under this license can be used under the same conditions as the CC-BY license except the data cannot be used for commercial purposes. More information on this license can be found on the Creative Commons website.

⁵ Bowser, A., Cooper, C., de Sherbinin, A., Wiggins, A., Brenton, P., Chuang, T.-R., Faustman, E., Haklay, M. (Muki) . and Meloche, M., 2020. Still in Need of Norms: The State of the Data in Citizen Science. Citizen Science: Theory and Practice, 5(1), p.18.DOI: https://doi.org/10.5334/cstp.303

⁶ Tulloch, A., Possingham, H. P., Joseph, L. N., Szabo, J. and Martin, T. G. (2013). Realising the full potential of citizen science monitoring programs. Biological Conservation, 165, 128-138. doi:https://doi.org/10.1016/j.biocon.2013.05.025

Back to top

To further promote the discoverability of published datasets, these can be shared via social media that may have a wide reach among passive data users and increase interest for newly published data⁷. Permission should be obtained from data providers to share information about specific datasets on social media. This can be incorporated into an existing archiving and publishing permission form provided by a data archive centre, where other elements such as data licenses and publication embargo periods are specified.

⁷ Rambonnet, L., Vink, S. C., Land-Zandstra, A. M. and Bosker, T. (2019). Making citizen science count: Best practices and challenges of citizen science projects on plastics in aquatic environments. Marine Pollution Bulletin, 145, 271-277. doi:https://doi.org/10.1016/j.marpolbul.2019.05.056

Case Study

DASSH shares news of newly published data via its Twitter page @DASSH.

iNaturalist share news of highlights and blogs about their data via social media account, @iNaturalist, which has a global reach.

Back to top

Once data are published, they can be made available on data portals, which enable the metadata and datasets to be findable and accessible. There are national and international portals, some of which are portals maintained by the data archive centres themselves. Examples of locations to find data, including those produced by citizen scientists, are:

• MEDIN Discovery Metadata Portal. A collection of metadata records for UK datasets archived by MEDIN DAC.. Users can search for metadata by title, unique identifier (GUID) dates, regions, season of interest, or specific controlled vocabulary parameters. The resulting metadata records will contain unique identifiers, contact details for data providers, download links, and general metadata about the dataset.
• DASSH Mapper. This shows all individual species records held in the DASSH database from standardised marine biodiversity datasets within the UK submitted by data providers. Users may search for specific taxon names, survey name, biotope, originator, owner, and they can filter for eDNA-derived records. Search results are shown in the map, and as a list further down the page where each survey should contain a link to the metadata for the survey from which the records was obtained.
• NBN Atlas. An online tool for accessing publicly available biodiversity data from the UK. Users can search for taxa or surveys and the results can be viewed as records on a map, or as charts summarising information. Metadata are given for these data, with links to the original datasets available.
• GBIF. International network and data infrastructure funded by the world's governments and aimed at providing anyone, anywhere, open access to data about all types of life on Earth. Similar to NBN Atlas, when searching for records the results can be visualised in a range of different ways, and the metadata and reference to the original dataset are also provided.
• OBIS. Global open-access data and information clearing-house on marine biodiversity for science, conservation and sustainable development. Data can be searched for and viewed on the mapper by filtering the results and saving the filters as layers. These layers of data can then be downloaded. The OBIS API can also be used to query the OBIS database to return specific information.
• EMODnet. Portals for data products in Europe are divided into 7 themes: Bathymetry, Biology, Chemistry, Geology, Human Activities, Physics and Seabed habitats.
• EurOBIS. Publishes distribution data on marine species, collected within European marine waters or collected by European researchers outside European marine waters. These data are also available on the EMODnet portal, and on a regular basis, all the EurOBIS data are sent to OBIS, which in turn sends its data to GBIF.
• Marine Recorder. Data from Marine Recorder can be viewed as a downloadable snapshot in the format of an Microsoft Access database.

Back to top

Metadata is essential for data re-use, as it provides important information about their provenance i.e. the materials and methods of data collection and any processing conducted and conditions for re-use. The metadata also prevents misinterpretation of data through the use of controlled vocabularies with explicit definitions.

Back to top

When using published data for further analyses, to inform research, reports, conservation and creating new publications from the data, the data license terms and re-use conditions must be followed, for example by crediting data owners, and not using these data for commercial purposes. Commonly used data licenses are:

• Open Government License (OGL). This license allows you to:
  • Copy, publish, distribute and transmit the data
  • Adapt the information
  • Exploit the data commercially and non-commercially
    However, you must:
  • Acknowledge the data source and include any attribution statement specified by the Data Provider(s) and, where possible, provide a link to the license information.
  • More information on this license can be found on the National Archives website.
• Creative Commons No rights reserved license (CC0). Under this license, the data provider has waived any rights to the data under copyright or Intellectual Property Rights law. This license allows you to:
  • Copy, modify and distribute the data, including for commercial purposes, without seeking permission from the data provider
  • More information on CC0 can be found on the Creative Commons website
• Creative Commons license with attribution (CC-BY). This license allows you to:
  • Share – copy and redistribute the data in any medium or format
  • Adapt – remix, transform and build upon the data
  • Data can be used for any purpose, including commercial use
    You must:
  • Give appropriate credit to the data provider
  • Provide a link to the license
  • Indicate what, if any, changes have been made
  • More information on this license can be found on the Creative Commons website
• Creative Commons with attribution non-commercial (CC-BY-NC). Data under this license can be used under the same conditions as the CC-BY license except the data cannot be used for commercial purposes. More information on this license can be found on the Creative Commons website.

Back to top

Re-using data to inform new projects, and to determine where gaps in existing data are, can make a project more sustainable as it can prevent the duplication of data collection events, it can save time that may be otherwise spent organising and running new data collection events, and can improve the design of a project if it is well-informed by existing data. By following FAIR principles and making data openly accessible to be reused, there can be further financial and environmental benefits. These benefits include:

1. Time saved by reducing time spent searching for existing marine data, as it should be easily and openly accessible. There is also a time saving benefit to organisations as their own data will be managed efficiently via a data archive centre.
2. Prevention of duplication of primary data gathering and research efforts as there is access to data collected by others.
3. Improved decision-making due to greater availability of data, as more existing data can be gathered, and better primary data can be collected.

Case Study

The MEDIN Cost Benefit Analysis report calculated a net benefit of £52.4 million over a 10-year period, from the services that MEDIN provides to its users. These include the accreditation of data archive centres, the MEDIN Portal for finding UK marine data, and a metadata standard to provide information about datasets.

Back to top