Social Strength

Understanding Real-world Friendship Strengths

This project is currently halted due to the Power being taken away from the People by the System (Jan 2015). More info here.

We at the ACSA Research Group want to create a breakthrough in the field of Social Networks Analysis by proposing a new and reliable way to obtain weighted social connections. In light of the current incentives for studying social networks with weights (on their edges) we believe to have come up with a simple and innovative solution to add weights to unweighted Facebook friendship graphs. These weights represent actual friendship strengths and can be interpreted in many ways, like trust between people, influence, or tolerance. It can also highlight socio-psychological matching, as well as show patterns in preferential communication between people.

While the methods and algorithms behind are under research, and thus secret, we need your help to validate out study on a larger scale. To help us, all you need is two sessions of roughly 5 minutes to run a few non-automated steps.

Disclaimer. We know that the data handled by our application is highly personal for you, so we will not ask you to upload anything on our servers. All the work will be done on your personal computer and we will only request you to send us the resulting graph file. That file contains the list of your friends (represented as a graph), and nothing more. The reason for privacy is that you will be required to run a java application that analyzes your Facebook private information history. Based on the found distributions, it adds numerical data - that quantifies friendship strength - to the final graph file. That's all. Your privacy is our concern as well!

To help us in our study please follow these steps:

Part I (3 minutes)

1. Log in to your Facebook account and click on the options menu (down-arrow) to the right of your notifications bar.

2. Go to settings.

3. At the bottom of your General Account Settings page you will see a link to Download a copy of your Facebook data. Please click it.

4. On the next page click Download Archive.

5. Then input your password,

6. And click Submit.

7. In the mean time please download our java application from here or cs.upt.ro/~alext/fb/fbapp.jar and save it to a known location (e.g. Downloads, My Documents, Desktop, tmp)

After this step you will receive an email with a zip archive with all your personal Facebook data. (PS. Good way to backup some photos if you didn't know that). The email will take a couple of minutes arrive.

Part II (5 minutes)

1. On your Facebook page, search for netvizz, or simply go to this link.

2. Once landed on the netvizz (v1.03) page, you will find a link named "personal network ..." in the upper half of the page. Please click it.

3. Then again, click on start (ignore the existing checkbox). This will retrieve your friendship network from Facebook, and will require a few minutes.

4. Once ready, you should be able to download your gdf file that contains the friendship graph. Right-click > save-as >

5. Give the file a name (e.g. your Facebook name) and save it to the same location as your Facebook archive and fbapp.

Once you receive the zip archive from Facebook and have your gdf file you can proceed to part 3.

Part III (2 minutes)

1. Download the zip attachment received from Facebook (~100 MB)

2. Open our downloaded java application named fbapp.jar by double clicking it.

3. Click the first Browse button and select the gdf file you have obtained from running netvizz.

4. Then click the second Browse button and select the zip archive of your Facebook account.

5. Finally, click on Generate and you should see a pop-up windows appear.

6. The pop-up asks you to save the resulting gdf file on your computer.

7. Please name the file using your exact Facebook name (e.g. John Doe) and save it to a known location.

Part IV (1 minute)

• Once you have obtained the final gdf file generated by our fbapp send us the resulting file (e.g. John Doe.gdf, of ~500 KB) via email to alext [at] cs [.] upt [.] ro

• Anyone interested in their own results can request some detailed analysis in their sent email.

Part V (5 minutes)

• Once you have submitted your gdf file, please complete this online form. We need to see how well your perception of your online friends matches the results of our algorithm. For this, we need the names of your best friends with whom you socialize the most. All the data from the questionnaire is kept private.

Thank you very much!

ACSA Team

cs.upt.ro/~alext/fb