# How Alato Lab Works

HOW ALATO LAB WORKS

Alato Lab provides automated detection of NZ bat species in recordings from
Alato AR4 /AR5 acoustic monitors. Upload recordings into Alato Lab and our
NZ-made AI model automatically detects the species in minutes or hours instead
of days or weeks.




Detections are categorised as Long-tailed, Short-tailed, Possible LT, Possible
ST or Non-bat. It is uncommon for Alato Lab to miss a bat as it is calibrated to
detect all ‘possible bats’ - there should be very few false negatives
(Non-bats), but you can expect 15-20% of detections to be ‘Possible bats'. We
recommend you manually review all ‘Possible’ detections. You can verify or
modify the AI detection by clicking the labelled buttons.

SIGN UP [https://accounts.alato.co.nz/sign-up] or SIGN IN
[https://accounts.alato.co.nz/sign-in] to Alato Lab
‍


CREATING SURVEYS

A survey in Alato Lab presents a period of time (e.g. 14 nights) where 1 or more
acoustic recorders were out in the field. A survey usually contains recordings
from multiple device locations. Device locations are automatically created when
uploading a survey folder with multiple device folders inside it.

‍




UPLOADING RECORDINGS

Alato Lab is optimised for uploading recordings from multiple devices at once
into a survey.

Uploading recordings from multiple devices

Upload the folder containing all folders of device recordings for your survey.
Alato Lab will use the second-level folder names (e.g. Device01) as device
location names for that survey. You can change device locations names after
uploading if you need to.



Upload recordings from multiple devices to Alato Lab.


‍



View recording devices in Alato Lab

Once recordings are uploaded, Alato Lab automatically starts analysing
recordings for bat activity. This can take some time, especially for large sets
of recordings. Once recordings have uploaded, you can upload more recordings
while others are processing.

‍



You can safely navigate away from Alato Lab while AI analysis is in progress -
you’ll get an email when analysis is finished.

‍

Uploading files from a single recording device

If you upload a folder from a single recording device, Alato Lab will ignore
second-level folders labelled with dates and use the top level folder name as
the recording device. If you have not renamed these and more than one has the
same name (e.g. ‘Bat’), Alato Lab will append this to ensure each device
location has a unique name.



Uploading recordings from a single acoustic recording device



Acoustic recording devices renamed in Alato Lab where names are not unique

‍

File processing quotas

Freemium accounts have a file processing quota of 5000 files (across all
surveys). If you need to process more files, upgrade your usage plan
[https://www.alato.co.nz/alato-lab#pricing].

Recordings already reviewed in Bat Search?

If you have previously reviewed recordings in Bat Search or AVIANZ, you can
still upload the same survey folder to Alato Lab - the bitmap files remain
unchanged, and Alato Lab will ignore unneeded files.

‍


BAT ACTIVITY REPORTS

After Alato Lab has finished analysing uploaded files you'll see an overview of
bat activity detected by Alato Lab's AI model. Filter activity by Species,
Device or Date/Time. Click any heatmap square for a filtered view of that
activity in Detections.

All metrics (such as Mean Detections Per Night) use calculations from the NZ
Department of Conservation best practice guidelines for bat monitoring
[https://www.doc.govt.nz/nature/native-animals/bats-pekapeka/resources-for-bat-workers/].
Heatmap detections are displayed in columns by ‘Recording night’, which is the
date that the device recording protocol started, including recordings collected
after midnight on the following day.
‍



Alato Lab automated bat detection report - activity by recording night and hour



REVIEWING BAT DETECTIONS

After Alato Lab has analysed your recordings, they are grouped into categories;
Long-tailed, Short-tailed, Possible LT, Possible ST or Non-bat. Use the labelled
tabs to filter by detection category.

It is uncommon for Alato Lab to miss a bat, but the AI model is calibrated to
show all ‘possible bats’, so you can expect around 15% of detections to need
reviewing. We recommend you manually review the 'Possible bat' and 'Non-bat'
categories.
‍



Automated bat detections in Alato Lab - review by category and manually verify
or modify

Verify a detection by clicking the button showing the suggested species
detection (dotted outline), or click a different button to modify the species
detection. Use keyboard shortcuts for speedy review.

When you manually verify or modify a species detection, that recording will move
to the bottom of the list in category tab it now belongs to. You can filter
detections by Review status to view only Verified or Modified detections. The
review status for each recording is shown in the .CSV export in Reports.

‍


GENERATING REPORTS

Once you have reviewed Detections, you can view summaries of the activity in
Reports and export the detection data.  The .CSV export lists all recordings in
the survey, with the most up to date ‘Species detection’ and ‘Review status’ for
each. The raw data includes both a ‘captured_at’ date and time (the ‘actual’
date the activity was recorded) as well as a ‘recording_night’ date, to help you
create reports by recording night.



Alato automated bat detection data export showing different 'captured_at' and
'recording_night' dates


All metrics (such as Mean Detections Per Night) use calculations from the NZ
Department of Conservation best practice guidelines for bat monitoring.
[https://www.doc.govt.nz/nature/native-animals/bats-pekapeka/resources-for-bat-workers/]

If you need any help getting started or have questions or feedback, contact
support@alato.co.nz [support@alato.co.nz]