what data is sent to GraphLab, Inc?

User 255 | 4/28/2014, 9:17:40 AM

While installing GraphLab, i noticed this message:

"GraphLab Create will send usage metrics to GraphLab, Inc. (when you import the graphlab module) to help us make GraphLab Create better. If you would rather these metrics are not collected, please remove GraphLab Create from your system."

what kind of logs graphlab sends to GraphLab, Inc? Also, if i use a custom dataset with graphlab create, does the GraphLab Inc. collects it too?

Regards Suvir


User 279 | 4/29/2014, 10:55:38 PM

Thanks for your note Suvir and for using GraphLab Create. We at GraphLab have strong values around transparency of data handling practices and security so we encourage you to read our detailed EULA which expands on the message you saw during installation: http://graphlab.com/legal/graphlab-create-eula.html

Also, should you choose to use a custom dataset with our product, we will not retain this information. Our information gathering is primarily for the purposes of improving our product and customer experience.

We hope yours has been a good one. Don't hesitate to contact us with further questions.

User 911 | 11/5/2014, 10:53:21 PM

Just as a note, the EULA no longer discusses what data is collected and kept.

<blockquote class="Quote">9. Product Instrumentation and Metrics. The GraphLab Create™ Product will periodically send usage data and metrics back to GraphLab for the purposes of improving the GraphLab Create™ Product, bug fixing, and crash reporting. GraphLab retains all right, title, and interest to this usage data and metrics (excluding any personally identifying information except to the extent aggregated and anonymized).</blockquote>

User 10 | 11/5/2014, 11:32:38 PM

Hey Suvir -

Thanks for reaching out for further clarification. GraphLab Create does not capture or record any dataset used with the product. Instead, we record usage metrics for the APIs (things like: did you create an SFrame, was the data loaded from s3, did you launch an EC2 instance, in what region was the EC2 instance launched). This usage data helps us understand what features are most utilized, which features are not being used, etc. This helps us prioritize new features, more documentation, bug fixes, and more.

Also, just to be very clear, we do not send the log files generated by GraphLab Create. They stay on your machine and are there to assist with debugging. Those are not sent as usage data.

Hopefully this helps clarify things further, please let us know if you have more questions.