Databrary Guide - Frequently Asked Questions

I qualify as an Authorized Organizational Representative (AOR). Can I sign as both Authorized Investigator and AOR?

No. If you qualify as an Authorized Organizational Representative you will need to have the staff member above you on an organizational chart sign for you.

Can individuals not associated with an institution be authorized?

No. To become an Authorized Investigator on Databrary you/your institution must meet All of the following conditions:

Have PI status or you are eligible to conduct independent research at this Institution and the Institution accepts responsibility for its Authorized Investigators’ actions related to the use of Databrary
Have all current research ethics training that addresses human subjects policy and issues required by your institution; and
Your institution must maintain an ethics or Institutional Review Board with U.S.-equivalent standards that reviews and approves research involving human subjects,
Be part of an institution that grants you the authority to choose and manage affiliates who meet Databrary’s qualifications and who agree to follow Databrary’s ethical principles, and
The person executing this Access Agreement on behalf of the Institution has the authority to do so
Data on Databrary is provided for non-commercial and non-personal use only and is subject to the terms of use outlined in the Databrary Access Agreement.

How much does Databrary cost?

For the time being, there is no cost to use Databrary.

However, consistent with the practices of peer data repositories and data services (ICSPR, Data Dryad, FigShare) we are developing a fee-for-service and subscription model that we will implement in late 2024 or early 2025. We encourage researchers with large datasets who are writing funding proposals to contact us in advance so that suitable data curation and storage fees can be included in the grant budget.

How is Databrary funded?

Databrary is supported through grants from federal agencies, and private foundations.

Where is Databrary located?

Databrary is located at New York University (NYU). All data is stored on servers that are maintained by NYU IT staff.

I did all registration and authorization steps on Databrary. I need to download a data set. My supervising professor has registered and gave me access but I still can’t download it.

Your supervising professor did register and requested access to Databrary. Access is not immediate. Your supervising professor and an authorized signatory on behalf of the university must review and approve the agreement. After Databrary receives the completed agreement, we may grant Databrary authorization to your supervising professor. The time it takes for this to be completed depends on the internal requirements of each institution.

Does Databrary offer “blind review”?

Databrary is a restricted access data repository that promotes the storage and sharing of identifiable and sensitive data. As such, Databrary has unique policies to protect the privacy of research participants. These policies govern who can access identifiable data and under what circumstances. The policies are codified in a formal access agreement that binds NYU, Databrary’s parent institution, and similar institutions where Databrary’s users conduct research. For the forseeable future, these policies make it impossible to support full double-blinded anonymous review of data and materials.

While some data and materials are shared with the public, Databrary restricts access to sensitive and identifiable information to people who have received authorization from an institution or to those whose access is supervised by an authorized investigator. Only authorized investigators can create and share datasets. Dataset creators, not Databrary, determine what level of sharing is appropriate for individual data files and for datasets as a whole.

So, the level of access a reviewer can have to a dataset depends on two things: Has the dataset been shared with the Databrary community? Does the reviewer have authorization from an institution to view restricted data or materials?

If both answers are yes, the reviewer may fully vet a shared dataset. Databrary does not report on access to specific datasets, so the reviewer can remain anonymous. Databrary does not mask the identities of dataset owners, so the blinding is only one-way.

If the dataset is shared, but the reviewer does not have authorization from an institution, the reviewer can confirm that a shared dataset exists, that the dataset has a persistent identifier, and that the dataset has a specific number of shared restricted access files. The reviewer may also examine any items that are shared with the public, stimuli, protocol, aggregated and de-identified datasets, etc.

What if an authorized investigator is leaving their institution for another institution but should still retain authorized investigator status?

The authorized investigator would need to meet all criteria required for authorization at both institutions.

How do I fix an audio file that is not playing in Databrary?

Download original session from Databrary and unzip the file
Open the free audio editing program Audacity
Open the file.
Select the downloaded file from the appropriate folder, usually the ownloads folder

Go to File > export > export as MP3
Use the following default parameters

If your file is named incorrectly (e.g. spaces/brackets/filename truncated) please edit the filename here.

A window pops up called ‘edit metadata tags’. just select OK
This new file needs to be replaced in the Databrary session folder.

How do I fix a video file that is not streaming in the Databrary application?

If you ever find that there are videos that don’t play in Databrary, the following instructions may be helpful to you.

use Handbrake (https://handbrake.fr/) to convert/transcode these files to .mp4 (this will change and fix the file if there is any corrupted block)
Go back to the session
highlight the video
click on the pencil in the upper left corner to edit
click on ‘replace’ in the upper right corner

The recommended settings in Handbrake that duplicate the .mp4 codec that Databrary uses during the transcoding process after a video file is uploaded are below. These settings may also be downloaded as a .json file and imported into Handbrake

Summary	Dimensions	Filters	Video	Audio
Format = MP4	Cropping = None	Deinterlace = Decomb	Video Encoder = H.264 (x264)	Codec = AAC (CoreAudio)
Select “passthru” metadata	Resolution Limit = 1080p HD		Framerate (FPS) = Same as source	Mixdown = 2ch Stereo audio (force to 2ch mixdown)
Align A/V Start	Anamorphic = Automatic		Variable Framerate
H.264 (x264) VFR video	Select ‘Optimal Size’		Quality = Constant Quality

Oops, I uploaded the wrong file in my session. Can I replace it?

Yes! Please follow these instructions:

Login to Databrary
Click on ‘Your Profile’ in the center of the screen
Navigate to the appropriate volume
Then the appropriate session
Single Click on the file that needs to be replaced.
Click on the Edit Button (the small pencil)
Ensure the appropriate file is highlighted in gray.
Click ‘Replace File’ at the top right of the screen
Navigate in the files pane to the correct file then select Open

I just uploaded a video file onto Databrary and it says my uploading is complete. Why can’t I see it on or download it from Databrary right away?

When you upload a video or audio file, Databrary converts it (transcodes it) into a common format that should be accessible well into the future. The original file format is stored in parallel. Transcoding videos takes time. Databrary will not show a thumbnail of your file until the transcoding finishes. If your file is especially large, then the wait time could be several hours.

If your file takes longer than that, contact Databrary support. Sometimes, transcoding jobs fail, and we can fix the problem at our end.

I downloaded files from Databrary, but the filenames are truncated.

When you download files from the Databrary application, the system appends information about the volume and session the files came from. This is to help you understand what you’ve downloaded. However, when Databrary downloads files there is a 32 character limit to the file name. So, if your file on Databrary has a long name, some of that name may be truncated in the downloaded file. Or, if the file name is too long (the first 32 characters of the filename are exactly the same as another file), it is impossible to distinguish between these files when they download and only one file will be downloaded.

There are two solutions:

Try not to create excessively long file names for files stored on Databrary.
Alternatively, you may want to download and use the databraryr R package. The package includes commands to download files from Databrary with the original (full length) Databrary file names intact. See this section of the databraryr guide for more information.

I have a dataset that has been completely collected. Now we are ready to share with Databrary. What do I do?

Send an Email to Databrary so they may assist you. You will need to include answers to the following questions:

What is the name and email address for PI for this project?
Do you have permission to share these data with the Databrary community?
- Was a Data Sharing Release or something similar completed at the time of Data Collection?
- Do you have permission from your IRB or Research Ethics Board used at your institution to share these data with Databrary?
How many files are there for this dataset?
What file types do you have (e.g. .mov, .mp4, .mp3, .pdf, .xls)? Databrary Supported File Formats
What is the total size of dataset (X GB)?
How many different sessions/participants do you have?
Do you have materials to share (e.g. data collection instructions, survey questions) in addition to the videos collected for the study?

Discuss this project with the Databrary Team to determine the next steps.

Format your data with the same file structure for all sessions/participants

Note

insert picture examples

Share data folder with the Databrary Support Specialist on Box.
Create list of files using Terminal (how to find terminal on Windows and Mac)
1. find [path to folder with files] > out.txt
2. E.g. find /Users/username/Box/Data_for_Databrary > out.txt
Download ingest_template.xls by opening the link, then click on the filename and select download.
1. Details for what needs to be completed on these templates are found here
  1. Make sure the items in the required columns are completed for each file.
    - the table Session Metadata describes the fields in the sessions_template.csv
    - the table Participant Metadata describes the fields in the participants_template.csv
  2. The participantID on the sessions_template and participants_template must match
  3. key is typically the same as participantID
2. It may be easier to complete ingest_template.xlsx then save each tab as a .csv
  1. participants_template.csv
  2. sessions_template.csv

I have a multi-site study. Is there a recommendation on volume/file naming schemes?

Try naming the each Volume: ProjectName_SITEID

The Short Name is INTERNAL only to volume Investigators and Collaborators

Where each SITEID has a unique id to each site, typically with the same number of characters.

Each file uploaded can have a standard naming scheme which includes the SITEID.

In the Volume Description, list the name of the university/institution in full. This allows for use of an API pull to ingest data when necessary.

After all data are collected, a new volume with ALL of the final data, including which data are from which site, can be created.

Is there a storage capacity limit for each project?

At the moment, no.

When we implement our fee-for-service/subscription model later in 2024, we will likely impose project-, researcher-, and institution- level limits for new projects that do not pay or subscribe.

Is there a maximum file size? We have some multigig videos and want to check ahead of time if there is a limit

At the moment, no.

However, we do not guarantee that large videos will transcode or stream well.

If you need to download videos en masse or large videos, please consider using the databraryr package. It can be much faster than the web application.

Note

We are still drafting responses to the following questions:

I have a multisite study. How do I add my co-investigators in the right order to the Databrary volume?

I am an authorized investigator. What is the best practice for revoking access to affiliates who originally were in my lab but have left?

Regarding permissions for users. We anticipate that we will have a range of users including faculty, professional developers, graduate students. We think getting access will be pretty straightforward for individuals connected to a university. However, how would someone (e.g., professional developer in a school district) gain access?