Donate your data

Late last month, Apple responded to the revelation that its iPhone and iPad products track their users' movements across mobile phone networks and WiFi networks around the globe.

Two British researchers had previously described at an April tech conference in California how this tracking data was being stored in an unencrypted file on the phone itself. They also wrote a data visualization program so that any iPhone owner could diagram their data on to an easy-to-understand map.

In a statement published to Apple's website, the company said users were "confused" about what exactly the company was doing with this data, adding that it had never tracked anyone.

However, in response to all of this, two German data visualization specialists are now asking for volunteers to donate their iPhone data as a curious project, known as Crowdflow.net, to map what these iPhones actually know. To learn more, Deutsche Welle spoke with one of the project's founders, Michael Kreil.

Deutsche Welle: What are you trying to do here? What's the ultimate goal?

Michael Kreil: The funny thing is that we don't have any goals. Because it started somehow with a scientific project. It started first when I analyzed my own iPhone tracking database. And I made a small heat map of Germany and all the places I've been to. And someone else on Twitter posted his database and so I took his data and visualized it too as a heat map of Germany. Then I had the idea that it would be interesting to take these two databases - his and mine - and to measure for example, at which time and which places we've been quite close to each other.

For example, last summer, we've been just 400 meters, roughly, apart at a demonstration, and then at a conference, we were 700 meters apart. Then I had the idea -- what would it look like to combine thousands of such databases and compare them, and what information is really in there?

It sounds like that at this stage, it's mostly curiosity more than anything else.

Yes, it's just scientific curiosity more than anything else.

Now, you worked on the data visualization with German politician Malte Spitz, right?

Correct, yes.

I interviewed Mr. Spitz about this visualization a few weeks ago. And that was such a great visualization. Is that what you're hoping to do, the same interface, or are you thinking about doing something different?

The idea behind data journalism is take the data, look into it, and try to find new relations and new knowledge, and then publish it. We don't know if we will publish the data in a web application, in a way that we did for the data retention from Malte Spitz.

Right now I'm looking at your blog, and it says this is from May 1, and it says ‘First database dump of cell and WiFi stations,' and there's a darkened map of the globe, and there's a bright spot in Germany and Europe, and there's a few in the US, and there's some in South Asia and one in Australia. What's been the distribution of data that you have so far?

Currently we have roughly 700 iPhone logs, and they're still growing. We hope that we collect more - perhaps thousands of such log files. And currently we have a lot of data from Germany, as this project started here, and we have a lot from Berlin. But the database is still growing and hopefully we will cover a lot of cell stations and WiFi stations in America and in other countries in Europe. And we also have a lot of data from Australia and Japan, and even India. We'll see.

It sounds like you don't even know where this data visualization project is going. Is that how a lot of data visualization projects go, that you learn something through the process?

When you combine a lot of data, you never know what kind of information is in data. Maybe you have one data set, which has a lot of information [but you don't know what it is], and you have a second data, also with no information, but at that point where you combine them, you can see what kind of information and correlations are there. That's the scientific part. The journalistic part is to combine and look and see if there is any knowledge or information there.

You can't say that ‘we want to prove that the mobile coverage is bad,' or that ‘Apple is tracking us.' You can't prove that because you don't know what's in there. You have to collect it, dig through it, and then you can see what's there.

So, in other words the story emerges from the data, and not the other way around?

Yep, definitely.

Now if I send you my data, is there any way to tell that this data came specifically from me, and specifically from my iPhone?

The correct answer is, I don't know. For example, [a few years ago] AOL released a bunch of data about search engines, and [it was later shown] that they could be de-anonymized. And [later] there was another example - Netflix.

They matched the Netflix database with the IMDB database. Then they can see that these people liked these movies and hated these movies. And then there's also accounts on IMDB with the same profile, so probably these people are the same. So they use the IMDB data to de-anonymize the Netflix data.

There's no way of telling if it's possible to use the track data to de-anonymize these people. I'm not sure, we don't know.

We started off that everybody can add their name to their log files. For example when we make a social visualization, that we can make some kind of a social structure. Currently we are saying that you can say that you're sending us the data anonymously, but we can't ensure your anonymity.

In other words, it sounds like what you're saying is that you're not sure if it can be fully de-anonymized or not.

We can't guarantee that it won't be de-anonymized.

But if I send you my file right now, is there any obvious way to know that it came from me, specifically, Cyrus Farivar?

We are deleting our logs and we're not tracking IP addresses and stuff. But you never know. Sony was hacked twice. If someone is hacking us and all this data is released, I just don't know. I try to make clear that we do all we can do to ensure anonymity.

Interview: Cyrus Farivar
Editor: Sarah Steffen