UK United Kingdom

Your social networks and the secret story of metadata

Researchers at MIT’s Media Lab have created an application called Immersion, which uses your email to display all of the people you communicate with in a highly visual way. Although it was designed primarily…

What does your email metadata reveal about you? Image from

Researchers at MIT’s Media Lab have created an application called Immersion, which uses your email to display all of the people you communicate with in a highly visual way. Although it was designed primarily as a way of illustrating a person’s connections and social networks, it has served to highlight the amount of information that is encoded in communication metadata and what can be done with that data without even needing the actual content of emails.

This is especially relevant at the moment with the revelations that the US secret service has been engaged in widespread surveillance using email and other personal data sourced from companies such as Google, Apple, Facebook and Microsoft.

Email metadata refers to information such as the sender and recipients of an email. On its own, this information may not be particularly very interesting, especially if as in my case, it is one of 53,129 emails that is in my work Gmail account.

But when visualised using a social network graph, relationships with particular people emerge along with their place in particular networks. Relationships emerge from the metadata that reflect working structures, projects I have worked on and companies that I have interacted with.

Author’s Email Network

The social network graph becomes even more revealing when combined with the email metadata from each connection in the social graph. Friends of friends become visible; connections between individuals’ social networks show how distant they are from each other.

You can see this visualised using another tool called the Challenger Network Graph that takes Facebook data and produces its own social network diagram. Unlike email, Facebook metadata contains much more detailed information about relationships between people. It is, after all, designed for this purpose. Consequently, the metadata is even more revealing than email in terms of showing the strengths of connections between individuals and groups.

Author’s Facebook Network

You can also visualise your social network graph on LinkedIn using an application called InMaps. As before, this graph is likely to highlight relationships from places of work, but will also show customers and other social connections.

Author’s LinkedIn Network

The aggregation of this metadata reveals a wealth of information about one’s personal relationships and highlights the power of having large amounts of small bits of information.

For the public, who are slowly becoming aware of the concepts of metadata and what it can reveal, this is an important step in understanding how it underlies our privacy on the internet.

For organisations and countries, the consequences of other countries or organisations having access to this data also comes to the surface. Even something as innocuous as social network analysis can reveal much information about corporate and political activity.

The beauty of this is that claims can be made that as long as you are not looking at the “content” of communications, you are not doing anything wrong. But one doesn’t need the content to reveal the full extent of what sort of communication is taking place.

Encrypting emails may make the senders and receivers believe that their communication is safe, but the mere fact that a communication is encrypted will make that communication stand out and highlight a peculiarity in that relationship. It is in and of itself another important bit of metadata that acts as telltale to the relationship and its attendant connections.

There is really not very much people can do to protect the digital traces left by metadata, short of not communicating electronically in the first place. Everything we do electronically leaks information through metadata, most of it hidden from its originators.

The fact that the Russian security services are resorting to manual typewriters, paper and presumably microfiche is a testament to the understanding that anything that is electronic is vulnerable to being leaked. Perhaps we will start to see a resurgence in the use of paper, pen and carrier pigeons.

Join the conversation

9 Comments sorted by

  1. George Michaelson


    Prof. Peter Linington from Kent explored email based social connection meshes about 15 years ago, and discussed it at DSTC during a research review of the CRC.

    I suspect that MIT's work here is serendipitous of the web 2.0 world where the metadata is available more widely: but the metadata *itself* such as the SMTP headers of mail messages, and trace logs held on Mail servers, has existed since before the RFC773 email standard was defined, and that was 1980.

  2. Michael Shand

    Software Tester

    Great Article, although the revelation from the 5 eyes is that they have the content of the email and the recording of the phone conversation

  3. James Hill

    Industrial Designer

    The diagrams look very much like those used by government agencies to solve crimes.

  4. Peter Hindrup


    I always knew that granddad with his quill and blotting paper had it about right!
    I would dearly like to see what 'they' made of my metadata!

  5. Pat Moore


    Interesting how these organic-looking clusters echo the shapes of brain system synapses or cosmic space bodies....electronically-woven tapestries, nets and webs. Or like electrified, interconnected cities on a map.. sucker burg cities...your face at the window looking out is his window in, electronic corralling. The metadata backend spider has his feelers on this globalised system of capture.....big brother who we've all come to know & love in mad Stockholm style....our 'dear leader', our global mind reader. As war systems technology it always was designed for such an application/app?

  6. Sam Han

    Lawyer; LLM student

    "The beauty of this is that claims can be made that as long as you are not looking at the “content” of communications, you are not doing anything wrong."

    Dr Glance, it might interest you to know that the US view on this is even more beautiful. Their law is that as long as you do not "query" the metadata, you are not doing anything wrong.

    Collection of huge masses of personal metadata are fine... as long as you don't look up any of it without lawful cause.

    Back on topic, those graphs are fantastic. Several years ago I read a fraud analytics paper which described such techniques, including how the FBI used them to find evidence in their Enron investigation.

    MIT researcher, Ethan Zuckerman wrote this great blog post on the same topic:

  7. Baneki Privacy Labs

    logged in via Twitter

    This isn't actually metadata; indeed, "metadata" has become an euphemism to cover for what's taking place. We follow Bruce Schneier in pointing this out (

    Metadata is "data about data." Metadata might be, for example, how many folks you connect with in an average 24 hour period. That's "meta." How many entries in a database table, or the name of the database table. Metadata. Data about data.

    This is traffic analysis. It's first-order data analytics…

    Read more