Researcher Platform

Researcher Platform FAQ

The following questions and answers pertain to accessing and using data within Researcher Platform.

I'm having a hard time accessing the Researcher Platform.

If you're having a hard time reaching the platform, please verify:

  • That you are either connected to OpenVPN or you have successfully logged in to WorkSpaces Secure Browser. See instructions within your product-specific documentation.

  • That you're using the correct URL that was emailed to you after you were approved as a Facebook Research Partner.

See Get help if you are still having difficulty.

What should I do when I get a "No Nodes Available" error when logging in?

Wait 10 minutes, then try logging in again. If the issue still persists, please feel free to open a case through Jira Service Management for further assistance. For more information and instructions, see Get help.

I encountered an error while attempting to access my environment.

Please try the following troubleshooting steps first. If you’re still experiencing an error after following these steps, go to Get help to ask a question. In this case, attaching a screenshot of the error message would be helpful. We’d be happy to investigate further.

Troubleshooting Steps:

  1. Make sure you’re accessing the environment from a compatible browser; we recommend using Google Chrome. For WorkSpaces Secure Browser, Google Chrome or Firefox is required.

  2. Use your browser Settings to clear all browser cookies and cache.

  3. If you are connected to OpenVPN, generate and use new OpenVPN credentials.

Why am I getting an OAuthException while running my query?

The error could be caused by an expired sign-in token. This happens when the browser is left open for too long. To resolve this, start by logging out. In JupyterHub in the top left corner, click File > Log Out. If you are connected to OpenVPN, generate new OpenVPN credentials (you can find OpenVPN instructions in your product-specific documentation). Then try to log in again. Please let us know if your error persists.

How can I use the data that I am approved to access?

Researchers may use data to advance research objectives that were documented in their access request.

What technical background do I need to utilize the Researcher Platform?

The Researcher Platform runs a modified version of Jupyter, an open source tool that supports multiple standard statistical packages, and provides a bridge to Facebook Graph APIs. Researchers gain access to a JupyterLab instance and can analyze available data in Python or R, including libraries like Pandas (for Python) or dplyr and gtools (for R).

Why don't I get consistent results when I run certain queries?

It is not unusual for this to happen in distributed data storage systems, as ongoing partitioning and indexing optimizations may alter the order of data serialization. If you need reproducible results (as opposed to samples), you must design your query in such a way that distinct results will be guaranteed. How you do this depends on the product and your objectives. Please reach out via Jira Service Management for recommendations.

Can I get access to a GPU for machine learning purposes?

Yes. Upon logging into Researcher Platform, you’ll be prompted to select a server option - either a notebook with CPU or a notebook with GPU (Tesla T4). For more information, see GPU Server Option.

Can I import other data into the Researcher Platform?

Yes. You can upload your own files to an Amazon S3 bucket, allowing you to leverage the data within JupyterHub. See Upload Files to an S3 Bucket for information and instructions.

Can I copy data from my Jupyter notebook and paste it outside of the Jupyter environment such as into my thesis?

Copying data and pasting it outside of the Jupyter environment is not allowed. However, you can export portions of your Jupyter notebooks. See Export notebooks for details about what exactly can and cannot be exported.

Can I install Python or R packages to a specific environment?

Yes. You can install Python packages available in the Python Package Index (PyPi) into your Jupyter environment using our Pip package manager class. You can install R packages into your Jupyter environment using our Conda or CRAN package manager class. For more information and instructions, see:

We generally avoid installing packages directly from GitHub as it is difficult to maintain and to support GitHub packages in the long run.