Social Data vs Sensor Data

Social data sensor data big dataThe two predominant kinds of big data are:

  • Social data and
  • Sensor data

Social data are data born in the social media realm such as facebook likes, linkedin updates, tweets and whatever the data entry we as humans do in the social sphere is called.

Sensor data are data captured by devices of many kinds such as radar, sonar, GPS unit, CCTV Camera, card reader and many more.

There’s a good term called “same same but different” and this term does also in my experience very well describe the two kinds of big data: The social data coming directly from a human hand and the sensor data born by a machine.

Of course there are humans involved with sensor data as well. It is humans who set up the devices and sometimes a human makes a mistake when doing so. Raw sensor data are often manipulated, filtered and censored by humans.

There is indeed data quality issues associated with both kinds of big data, but in slightly different ways. And you surely need to apply master data management (MDM) in order to make some sense of both social data and sensor data as examined in the post Big Data and Multi-Domain Master Data Management.

What is your experience: Is social data and sensor data just big data regardless of source? Is it same same but different? Or are social data and sensor data two separated data worlds just both being big?

Bookmark and Share

5 thoughts on “Social Data vs Sensor Data

  1. Vishu Rastogi (@decembers_child) 24th March 2013 / 07:53

    Hi Henrik,

    Both constitute Big Data. But would it be fare to compare them? I feel it would like comparing Apples and Oranges. The only thing common between them is that they are generated in huge volume, variety and velocity. Both Social data and Sensor have a rightful place, but i would dare to say that they do have an overlap where in, one can feed the other Eg: I could feed my facebook likes an RFID when ever i go shopping, but trying to replace one with other could result in serious inaccuracies.


  2. Henrik Liliendahl Sørensen 25th March 2013 / 05:30

    Thanks for joining Vishu.

    On Google+ Jesse Johnson says:

    It seems like with both social data and sensor data, there are two different sources of structure: The first is the network itself, i.e. the connections between people or the relations between sensors (such as physical proximity.) The second is the activity data: likes and tweets on social networks or the activity reported by sensors. In both cases, one needs to understand both structures and the relationships between them in order to get the full picture.

  3. Harsha Srivatsa 25th March 2013 / 09:54

    I work in both domains and would like to share my experiences with them. There seems to be similarities and differences depending on how you view them.

    I think that there needs to be a differentiated perspective on how M2M and H2M (Human to Machine) interactions needs to be organized.

    With human interactions we try to analyze sentiment and behavior and often the end result is to cater to human end user. With M2M do we really to cater back to the generators/end users? Apart from figuring out that machines are working (in some cases human generated events) do we really need to cater back to them?

    Sensor data from Internet-of-Things can be characterized with predictable velocity and data growth. Human generated social data growth is hard to characterize and predict.

    I also suppose that in the structured – unstructured data equation, sensor data tends to lean towards the structured side with somewhat simpler connectedness whereas social data is very much on the unstructured side with connectedness that often needs to be later derived and analyzed.

    Ultimately data is data is data and often ends up in the same data island. Do we really need to care where it came from? Or do we care more about getting it analyzed and get actionable results from large scale analysis?

  4. Henrik Liliendahl Sørensen 26th March 2013 / 17:40

    Thanks a lot for commenting Harsha.

  5. Dave Chamberlain 26th March 2013 / 21:57

    I think that we should consider systems/applications/infrastructure log data needs to be a separate source, although you could argue that as it is generated by software it falls into the same category as sensor data.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s