Jump to Section:



Background


Large-scale biodiversity data repositories may contain incorrect data, whether that be from accidentally misidentified species or from general mistakes in data entry.

The Atlas of Living Australia (ALA) uses automatic cleaning and filtering mechanisms during data ingestion. Though, because the intake of data at the ALA is often incredibly high – we are not resourced to manually assess every individual occurrence record received – and, therefore, there may still be errors.

 

As an aggregator, our function is to bring together many sources of biodiversity data for everyone to freely use. We are not the data custodians or owners. This means that when there is an issue with data – for it to be corrected, the data owner ultimately needs to make changes.

To help alleviate this issue, when you “flag” a record, the ALA allocates it to our automatic filtering system as suspicious. It is then removed from automatic data profile searches. A notification is also sent to the data owner to prompt them to make changes (if they see fit). Outside of erroneous data or data with mistakes, there may also be reasons to flag data for biosecurity reasons (see Table 1 below).

To maintain transparency, you can always see what has been updated in a record from our cleaning and filtering mechanisms or by a data owner by visiting the “compare ‘original vs processed’ values” region of an occurrence record page. If you want to view larger groups of possibly erroneous data that have been uploaded to the ALA, you can turn off a range of filters in the “data profile” section after selecting “view records” underneath the occurrence records map on a species page (also known as the “Biocache” search function).

If something seems incorrect with a record and you can reasonably justify it – let us and the data provider know with these steps.



When should I not flag a record in this way? 


For records that come from iNaturalist, it's best to go directly to the source, flagging the issue on the iNaturalist record itself. This is to ensure the original creator is notified of the issue. 


You can find the link to the iNaturalist record on the ALA occurrence record, in the Occurrence ID field.

An ALA record page showing the link to the iNaturalist record. it is in a table in the "Occurrence ID" row


Most often, the issues identified on iNaturalist records are species or taxon misidentification. To flag this on the iNaturalist record, scroll down to the Activity section of the page, and under the suggested IDs already listed, you can 'Suggest an Identification' by typing in the text field below, then clicking Done.

suggested species names from two different users, then a section below with a text box to suggest an identification



If the issue isn't a species misidentification, you can still add your feedback on the iNaturalist record. Scroll down to the Data Quality Assessment section. Here you can vote a thumbs up or thumbs down for the various categories, such as 'Location is accurate, or 'Date is accurate'. Voting doesn't generate a notification for the user, so it's good practice to also leave a comment (in the Activity section of the page) indicating you've voted, to give them a chance to respond and potentially fix the record.

the table of data quality options, with thumbs up (yes) and thumbs down (no) for each




How to flag a record: 



1. At the top left of an occurrence record page (use the interactive map or “view records” from the species page to select a single occurrence record) locate the “Flag an issue” button, click on it. If you are not signed into the ALA, you will be prompted to do so.

 

screenshot of a species occurrence record page with the 'flag an issue' button circled in the top left of the page



2. In the pop-up text box, start by choosing what the issue with the data is using the dropdown menu (see Table 1 below for descriptions of each). 


screenshot of the dropdown menu showing available types of issues with records (see table 1 for a full list)



Table 1: Issue types

Geospatial issueIs this record suspiciously out of its range?
E.g., Are the latitude and longitude in the correct positions? Have they been accidentally swapped?
 
Habitat incorrectLike the above, is this record located in a habitat that is incredibly unlikely?
 
Suspected outlierThe data are in a position outside of a general range that seems unlikely. This can skew results when using authoritatively.
 
Taxonomic issueThere is an issue with the species name or how it has been categorised in the taxonomic tree. Note that the ALA uses three authoritative sources as the base for our taxonomic tree: the Australian Faunal Directory (AFD)  f or animals, the Australian Plant Name Index (APNI) / Australian Plant Census (APC) for plants, and the New Zealand Organisms Register (NZOR) for relevant species of New Zealand. Where gaps exist in these trees, other naming authorities are used. See our article on taxonomy for a more comprehensive overview of how the ALA uses naming authorities. 

Taxon misidentified Can you tell from the images that this species has been misidentified as something else, perhaps similar, or completely different?

Temporal issueIs the date or time that this record was logged as occurring suspicious? Does the species usually emerge at this time and in this area? Is the date of the record incomplete (e.g., just recording the year, or suspiciously marked the first day and month of the year).
Duplicate record

  • Duplicate occurrence



  • Tissue sample

  • Split sample 
Is there another record that seems to be exactly the same individual as this one? Include the duplicate's record ID (e.g., 136903182).
  • There is another occurrence record with the exact same details in the ALA, or it can be identified as the same individual through images somehow (e.g. legible picture of an ID band on a bird with a unique number, or unique individual feature)

  • This occurrence record is of a tissue sample of an already lodged occurrence record

  • Identical specimen or individual sent to multiple herbaria with different record IDs; but the same coordinates, collector, and often locality description

Biosecurity issue Is this a record of an invasive species that has evidently migrated out of its general range, or is it a completely new sighting of the species in Australia? See the Australian Government list of Weeds of National Significance (WoNs). 
 
Other issue None of the above describe the issue adequately. 

 



3. 
In the comment space, write succinctly, but with as much detail as possible as to why this record is incorrect. You can refer to academic sources. You can check the box below if you would like to be notified of a change due to your flag.

 
screenshot of the comment space and submit button



If you have any further questions about flagging records, please reach out to us at support@ala.org.au.