5/26/2016

9 reasons why "Big Data", "IoT" and "Smart Home" might be so dangerous...

...or delicate at least


Not only buzzword paradise (for the big companies) - but big money

Data streams

Working for years on making "things" communicate with "machines" and hence putting them in the augmented live of ourselves there is always the discomfort of the consequences of one's actions.

I wouldn't dive any deeper into any horror scenario resulting of this or serve a model for another (american) dystopian movie.
Let us instead analyse hard facts in form of ONE single data stream only being available as a graph so that you can judge yourself what could be a realistic outcome of the "Internet of Things" where already billions and billions of devices are connected worldwide.

Day 18 since seed
As you can read in my blog I'm currently tinkering with the new IKEA Krydda indoor gardening stuff. Nature should be allowed to run its course? Sure, but adding electronics always makes fun and often gives much more insight to things you didn't see before (did you ever take a thermal picture of something? See examples here). I'm putting the data into the "Internet of things" here so you can see 24/7/365 what's happening in my gardening paradise at home.


Day 18 since seed
So let's see what we can tell just from one and only one graph ("data stream"). For the beginning we assume we even have no absolute numbers like YOU do it everyday 24/7/365 with your smartphone, PC, car etc. sending your life ("numbers") to the big data centers in Silicon Valley and everywhere in the world analyzed by the smartest algorithm programmers available and fastest electronic brains existing.

ONE graph


1. Flatline

Nothing happening and therefore not interesting while there is nothing to see in the graph aka as a "flatline"? Wrong! Nothing happening could mean you are sleeping (now "they" can tell your sleeping habits if there are more time periods to sample) or find out that the apartment is not occupied (welcome burglars!). I do not even need to know the type of data (like temperature or electric power consumption).

2. Controlled curve

There is something ramping up for a certain amount of time in a very constant way. First guess: Might be a (electronic) device not a human action. Maybe the heating system? Without knowing the data type (8) it's hard to tell. But if I only know once what type of device this kind of ramping-up produces, I could draw conclusions to every other data stream from anyone I get with the same pattern.
In this case it's the wake-up mode of a Philips Hue lamp over a time frame of thirty minutes (7). Linked to number (6) you might even tell the lamp type or color. Hackers welcome if they know which devices are in your home! And your next ads in your browser/ app might be related to the lamps. You also need a light strip from Philips?

3. Elevated flatline

The previous phase is completed. Might there be something happening next or what? Compare 1, 2 and 3 over more periods and you already get very interesting details on sleeping and living habits and hence patterns.


 

4. Digital

Something gets elevated very quickly like a 0 to 1 transition. Knowing the type of data (8) it's not hard to judge what happened. More insight into your living habits if linked to (7) time and date. Just from one (mechanical) action. Imagine all the mechanical switches are replaced with "intelligent" ones.

5. Analog

Something "irregular" is happening. Typical analog data of very fine granularity and therefore very helpful. Put the data in the big electronic brains and compare it to typical analog data (e.g. weather data like temperature, wind, cloudiness) and there will be a meaningful outcome - for sure. If you have the data type (6, "light intensity"), (7, "time"), (8, "light") you do a favor to the wallets of the big data center owners because they only need to buy little computing time. Now "they" not only know your sleeping habits, living pattern, your electronic devices but also where you live.
No they did not need your GPS signal you are always sending with your smartphone or the WiFi data you are allowing them to track.

6., 7. Goody: Scaling/ reference

We are still not talking data streams with exact (absolute) numbers but only a single graph we get over time. Add goodies like a scale or reference to make it easier. But be assured clever algorithms are already bored if you give them such kind of a cheat sheet.

8. Data type

Oh "they" also know what kind of data you are delivering - thank you for willingly helping to get the graph-only data even more interpretable.

9. Big data

You are adding more linked data like temperature, humidity and a data stream with numbers? You are already toast, this only adds to the finer granularity of your being. And always remember: you still have not told someone one single (unencrypted) letter - like you do everyday putting your most intimate details on messengers like "Whatsapp" or by sending your data into the "cloud".

Conclusion


We only examined ONE very simple data stream in form of a graph. Only from the pattern of one simple, single data source you can tell a lot of stuff. Much more then mentioned above - we only scratched the surface in interpreting data. That's why educating people as "Data Scientists" exploded in recent years.

Now imagine ALL the data YOU (and of course ME!) are sending every second into the enormous data cloud! Judge yourself what could happen and what we should do in the future regarding data aka the "Internet of Things". For the moment OTHER people are getting YOUR data and earning a lot of money from it or even worse.

What we should do and behave? Proposals very welcome! For the time being enjoy this "paradise" of "Big Data" - and did you just switch your light on?

No comments:

Post a Comment