-
synapfire.com
The term Big Data is commonly used to describe the ever
increasing abundance of information being gathered, stored, and
analysed in todays world. Whilst this explanation provides some
insight into the meaning behind the popular buzzword, Big Data is
truly defined by its characteristics, and what stands it apart from
historic data analytics.
DEFINING BIG DATA We often associate the 4 Vs with Big Data,
although more accurately this is really the 5 Vs, as we shall
see.
Volume refers to the vast quantity of data now being accumulated
each second, and stored for future use. It has been calculated that
the volume of data created since the beginning of time, up until
2008, will soon be generated every minute. This means Zettabytes of
data. A study at the beginning of 2014 noted that we are now
accumulating data at the rate of 2.4GB per second.
This is hard to comprehend, but a good visualisation is to
imagine a truck turning up at your house every hour and unloading
8640 1TB external backup drives into your home. This reoccurs each
hour, with none removed, and this allows some understanding of the
volume of data we refer to.
Velocity is the speed at which this data is created, mined and
moved around. Think social media, and in-memory analytics for rapid
interrogation and recall of information for strategic benefit.
Variety is what differentiates this data mining from standard
analytics. We are now not only storing text, but images, video, and
all manner of unstructured data. This data cannot be indexed
through standard methods and common database practice. A new
approach to index and recall is required to handle the unstructured
data.
Currently the volume split of data being created each day is
thought to equate to around 20% structured, or classic data we are
used to dealing with in tables, and
80% unstructured - images, video, social commentary, etc.
Veracity is the integrity of the data stored, and how it is
used. A key point to keep in mind when analysing data is that
social content is most often opinion and conjecture, not hard
unequivocal facts.
Value is the 5th V. Whenever we perform data analytics, the key
driver must be to generate, add, or present value in what we are
doing.
AUTOMOTIVE: VEHICLE ANALYTICS The automotive sector is one where
the implementation of sophisticated analytics was almost
inevitable. Whilst engineering practices have advanced, and the
technology within our cars has been steadily improving, there have
been no major advances over the past 100 years when it comes to
maintenance of the vehicle.
For the most part this still operates on a system where a driver
takes their car to the dealer when there is a suspected problem.
The dealer analyses the car, determines the problem, and then fixes
it. A computer now does a lot of analysis, but it is still
dependent on a system where a fault must first occur.
With technology advancing at the paces we are now seeing, this
is no longer good enough we as consumers demand more.
Automotive Analytics Driving in the Cloud
-
synapfire.com
VEHICLE DATA MONITORING Sensors are everywhere, and we encounter
them all the time, but are now so used to them that they blend into
our everyday life without direct impact or interference. They
monitor doors to allow automatic opening, they measure and control
air temperature, they save us energy through dimming and switching
off lighting when there are no occupants in a room. Sensors are
already embedded into our lives.
Within a typical modern car, there are hundreds of sensors, and
within some of the more advanced cars the sensors total thousands,
on par with that of early Boeing 737 aircraft.
What we must remember is that these sensors are, for the most
part, generating structured data. That is prescribed information
based upon a known state or expected or permitted outcome. For
example a sensor monitoring the health of a headlight bulb will
know that it is reporting whether the bulb is working or not. If it
is not working, it cannot report that the reason is because a stone
came through the glass and took out the bulb.
If you think of most components of a car, and use physics to
boil things down to their most base principles, the sensors are all
reporting known potential outcomes. This is even true of systems
that we are in awe of on recent luxury cars such as variable speed
based on traffic avoidance.
The sensors in this scenario are reporting current speed,
distance between vehicles, and closure rate, to make corrections to
the speed of our vehicle and maintain the safe gap.
When different sensor data is brought together, we achieve swarm
intelligence, as together they help to interpret scenarios and
events being experienced by the vehicle and its occupants in order
to understand causality and resultant impact.
So how can this data and the many thousands of sensors advance
the motoring industry, and how can the consumers experience in this
sector evolve into what is expected in a modern technology focussed
world?
We can now store this incredible volume of data, we can harness
it through advanced analytics, and when combined with the evolution
of telecommunications, 3G, 4G, 5G and so forth, we have a way of
letting others access the data.
This breaks the cycle of user reporting problem and opens the
doors to proactive intervention.
PRE-EMPTIVE MAINTENANCE With sensors collecting data on the wear
status of crucial components of the car such as brake pads, discs,
fuel filters, we can now proactively monitor the status of the
vehicle. This data is fed back to the dealership where steps can be
taken to notify the owner should wear reach the point that mandates
some action be taken.
Taking this to the next level, algorithms can monitor patterns
across vehicles to identify potential issues on a range, or model,
before actual problems present themselves. This is where we truly
start to see the power of data analytics. Engineers can analyse how
components wear and interact when placed in real driving
conditions.
Currently manufacturers perform thousands of miles of testing in
artificial environments, and on test tracks, but nothing can
reflect the true value of real world road experience spanning all
geographic locations within which these cars are to be sold.
The knowledge the manufacturers can build based on these data
feeds is extraordinary, and starting to become a reality now.
Several vendors are getting involved in this space, including BMW,
Audi, and Toyota.
The interesting wild card in the group is Tesla. With their
all-electric cars they have certain advantages and an edge over
their combustion engine counterparts. They are already starting to
capitalise on this through proactive overnight maintenance of their
vehicles at peoples homes.
To give an example, where the engineers are receiving data
streams about handling and motor response, they are able to gather
this data, simulate changes, and then if deemed necessary, push a
patch to the
-
synapfire.com
vehicles. Because the vehicles are connected by Wi-Fi to our
homes, subject to agreement, this means patches to fix and improve
our cars handling, performance, and economy, can be downloaded to
the vehicle overnight.
This is utilising Big Data analytics to improve a product that
is already in use, and is a good example of what the market is
referring to as Innofusion, or the melding of different innovative
technologies and scientific disciplines to harmoniously improve a
process, action or known experience.
But maintenance is not the only aspect of the automotive
industry to benefit from Big Data, as other areas start to realise
the power of the consumer crowd, and their voices in several areas
of the industry.
CONSUMER LED DESIGN This is an area where we are seeing an
explosion through sites such as Kickstarter and Indiegogo, with the
crowd dictating the design or end product. Coupled with the power
of data analytics, trawling unstructured data across forums and
social media sites, you can begin to create consumer targeted
manufacturing.
Audi have been doing this for some time now, and started with
crowd sourced input into the design of their audio systems within
their cars. When they update the in-car entertainment strategy,
they look to the forums and invite input into what is important to
the drivers of their cars when it comes to audio entertainment.
Audi have successfully done this now for several generations of
audio equipment.
But it doesnt end there. Many manufacturing industries are
constantly scouring social commentary and major forums for
information and views on their products, and the automotive sector
is starting to leverage this power. The idea of harnessing the
power of collective design is a very real and powerful marketing
and design tool, often heralded as The Crowd as the designer or
Crowd as the architect. Whilst this may not be 100% accurate, and
the degree of input taken in can vary from company to company, any
manufacturer of consumer based devices and products that does not
tap into this pool of ideas is orchestrating their own demise.
So with Big Data contributing to not only the design of our
future vehicles, but the pre-emptive maintenance of tomorrows cars,
we are beginning to see a more
seamless integration of services across the present automotive
sector - a synergy across all the required independent building
blocks that make up our daily experience as car owners and
drivers.
INTEGRATED SERVICES With the dealer networks now able to keep a
closer eye on the cars operations, and pre-empt potential problems
or risks, closer ties can be made to recovery operations, insurance
companies, and other important participants in the car ownership
chain. We are moving closer to the point where recovery trucks
could be en-route to your predicted breakdown location before the
breakdown occurs.
This sounds a little bit too much like the film Minority Report,
but why not? Imagine a time when a red light flashes on your
dashboard, the engine stops and you pull over on a country road.
Within a minute you receive a call to inform you that all is in
hand, and a recovery truck is already en-route with a firm ETA.
Back at the dealership, they have already run diagnostics on the
data, identified the failed components, and an order for the parts
is in progress.
You are picked up by the truck, and taken home. Meanwhile the
dealer contacts you to confirm that your car is en-route to the
dealer, and will be ready for collection by a provided data.
This sounds great, if not a little sci-fi, and potentially a
little big brother, but this is nearer to a reality than you may
think. Several components of this vision are already in place, with
manufacturers beginning to look at partnerships and closer ties
with insurance companies and repair organisations, to steadily move
us towards a more seamless ownership chain.
DATA SECURITY With all of this data being stored and recalled
for analysis, there must be assurances of anonymity for the users
providing this data. The crucial importance of lessons learned so
far cannot be forgotten.
How this anonymity is maintained and adhered to by the
corporations presents several dilemmas from commitment to
enforcement. This is an area of data gathering and analytics that
needs serious discussion, as current data protections legislation
is not fit for our rapidly accelerating and globally connected
world.
Big Data is an important part of our evolution, but care is
needed to ensure that data analytics does not betray our privacy
for corporate gains.