Big Data top down

Manageable by reduction

More than 90% of all available data was generated in recent years. This relates to the traces that we leave behind when we go online. Via Facebook, LinkedIn, Twitter, and others – as well as through our surfing and viewing habits, purchasing behaviour, use of media, travel details and so on.

So how does this stream of data become manageable? Datamining and mathematical algorithms reduce the data to movements and forms of behaviour that apply to a large group of people. There are applications in various areas: traffic jam issues, predicting flu epidemics, demonstrating spikes in purchasing habits, assessing safety problems, etc.

Context

However: collecting lots of data very quickly using computer systems ignores the essence of data-gathering: understanding why people behave in a certain way. Because Big Data focuses on the actual behaviour displayed, the context of the behaviour remains out of the picture to a large extent, which calls the practical usability of Big Data into question. Only by placing the context of the individual action at the centre can we deduce what the reasons were for a certain action taking place.

3 dimensions

The context of an activity takes shape through 3 dimensions:

1. The spatial dimension – the place where people are.
2. The social dimension – who was there; level of satisfaction; feelings of stress, time pressures; motivation, etc.
3. The temporal dimension – how long does an activity last (duration), when does the activity take place (timing), how many times is the activity repeated (tempo), what other activity precedes/ follows the activity (sequence)?

The more context that can be collected, the better we are able to colour in the context in which the behaviour takes place.

How to collect it?

We use the MOTUS software platform to collect information about people’s behaviour in 3 ways:

1. Retrospective registration of previous actions – limited to 24 hours in time, to a maximum of one week in certain situations.
2. Continuous registration for recording actions in real-time.

Both of these methods require an active effort on the part of the user. The design and flow of the MOTUS app encourage the recording of actions in (virtual) real-life.

DRIE!!! Passive registration via plugins, apps, wearables and sensors – these devices collect information about length of sleep, geolocation, stress levels, travel patterns, attendance, user behaviour, etc.

Additional contextual questions may be asked for each of the three forms of registration.

Through the 3 forms of time registration and the structure of the context, we arrive at Personalised Big Data.

The database established in this way make it possible to gain insights into the ‘why’ of people’s comings and goings. This ‘why’ question is essential for predicting future behaviour.