Thursday, April 26, 2018

BigData Analytics: Color Vision in a Cat Whisker World

Seeing with New Digital Eyes

"He who fights with monsters should look [into] ... himself ... if you gaze long into [the] abyss, the abyss also gazes into you." -- Friedrich Nietzsche (1844-1900), from Beyond Good and Evil -- Aphorism 146

Natural Analog Question: For jungle predators, what is the advantage of seeing in color -- when all your prey only see in black & white -- or maybe only have cat whiskers and no eyes?  

Now, thinking in metaphorical terms: What are the computational insight advantages of BigData Analytics -- and social media data curation & Data Mining -- over everyone else? Are you and your electronic Doppelganger food for BigData predators?  How can you detect stalkers?  Are there taste testing teeth marks on your electronic doppleganger?

Ross Anderson of Cambridge University offered up a form of these questions in his 2017 YouTube interview -- entitled "The Threat." Mr Anderson illustrated 30+ years of trends for potentially predatory BigData aggregation and exploitation -- and Mr Anderson's odd journey thru "technical gettos" -- political gaming of technology -- and vetting the social/ psychological/ economic landscape of "digital life". 

One glaring exploitable aspect in a world gone Gaga with "free" social media? Economic & Political manipulation of "digitally curated" social media relationships.  Too scary to believe? Hear Ross Anderson's perspectives:


2017: "The Treat" -- 30+ year trends in
in computing, privacy, social adaptions (or
not) to BigData Surveillance Capitalism

Deep Nerd Hints: You can download audio only of this interview -- and the audio contains 98% of the key info -- via the command line tool YouTube-DL.  See Nerd Bonus addition at end of this posting.

In 2013 -- long before the Facebook and Cambridge Analytica controversy arose, Maxine Waters hinted at the ice berg under the digital waters -- for naive sailors of the digital seas:


2013 Maxine Waters perspectives
on Database Nation & politics.

Uncharted Ocean? No. Surprise History and 17th Century BigData: Just before the 1789 French revolution, the King's secret police were "monitoring" the "news" of Paris.  How?  In a time when most folks were illiterate?  Street vendors would "sing the news" -- to customers and each other.  News traveled by song -- and by word-of-mouth -- and was a mixture of true events and "fake news" (salacious gossip).

The King's secret police would write up a Zeitgeist summary of this street "news" -- apparently for "actionable" intelligence.  Effective?  17th century BigData did not save the French King from the revolution. And illustrates how BigData machinations of economics & politics is not new. One amazing character and secret police monitor was Jean-Charles-Pierre Lenoir -- escaped the worst of the French revolution -- and left valuable historical insights into the approach and unfolding of the French Revolution -- from "loser side".

Not paranoid? Yet?  Check for teeth marks on your electronic doppleganger.  Some are as subtle as search engine stovepiping of your query results. Many are changes in online Ads boiling up in web page visits.  Tone and flavor of bulk mail is another indicator. And more regular & routine ID theft challenges. And more ID / authentication friction at post office, banks, credit-debit card use, good & bad police interaction.  Surveillance capitalism data trafficking -- coupled w/ rouge BigGov actors -- begets more ID theft, which begets more surveillance friction, which beget more ID theft, etc etc.

Upshot?  Action?  Lacking deep resources to thwart BigData with your own BigOther magic?  If you must be on the internet, then first know this: We cannot stop the data gathering -- only pollute it.  Or at least blow a smoke screen.  

Other Options?

(1) Grow some color vision -- learn to detect teeth marks on your electronic Doppleganger.  (2) Learn how to use some digital camouflage -- and -- metaphorically speaking -- learn to use the old Star Trek technique of "rotating your shield frequency" -- to throw off BigData alien tractor beams molesting your electronic doppleganger: Create "spare" online profiles.  (3) Use app browsers that thwart BigData tracking -- like Firefox Klar -- or on WinXX machines routinely flush Google Chrome browser cache (amazing how your search results change!) -- avoid "logging in" -- and (4) Boot and cruise internet with "stock" live systems.  

Welcome to a Brave New World, Mr Savage (;))!

Nerd Bonus & Hints: How to fetch audio only track from YouTube videos:

From your *NIX command line: 

# =============================================
# --- Fetch audio track from YouTube URL
# --- install YouTube-DL if not already onboard
sudo curl -L \
  https://yt-dl.org/downloads/latest/youtube-dl \
  -o /usr/local/bin/youtube-dl 
sudo chmod a+rx /usr/local/bin/youtube-dl 

#
# --- Query YT for stored stream containers,
# --- "shopping list" output results below.
# --- Note selection of container 171 for audio
/usr/local/bin/youtube-dl -F rSw0fDCSfX8

#
# --- Fetch audio only track:
/usr/local/bin/youtube-dl -f 171 rSw0fDCSfX8

[youtube] rSw0fDCSfX8: Downloading webpage
[youtube] rSw0fDCSfX8: Downloading video info webpage
[youtube] rSw0fDCSfX8: Extracting video information
[download] Destination: The Threat by Ross Anderson-rSw0fDCSfX8.webm
[download] 100% of 26.35MiB in 00:36


# ===============================================
# --- Typical output: Query YouTube for shopping
# --- list of video container IDs and digital 
# --- formats -- for help, use "-h" with youtube-dl
/usr/local/bin/youtube-dl -F rSw0fDCSfX8

[youtube] rSw0fDCSfX8: Downloading webpage
[youtube] rSw0fDCSfX8: Downloading video info webpage
[youtube] rSw0fDCSfX8: Extracting video information
[info] Available formats for rSw0fDCSfX8:
format code  extension  resolution note
249          webm       audio only DASH audio   52k , opus @ 50k, 15.80MiB
250          webm       audio only DASH audio   72k , opus @ 70k, 19.07MiB
171          webm       audio only DASH audio   89k , vorbis@128k, 26.35MiB
251          webm       audio only DASH audio  129k , opus @160k, 34.69MiB
140          m4a        audio only DASH audio  130k , m4a_dash container, mp4a.40.2@128k, 39.80MiB
278          webm       248x144    144p  105k , webm container, vp9, 30fps, video only, 29.10MiB
160          mp4        248x144    144p  107k , avc1.4d400c, 30fps, video only, 15.71MiB
242          webm       414x240    240p  242k , vp9, 30fps, video only, 49.02MiB
133          mp4        414x240    240p  258k , avc1.4d400d, 30fps, video only, 45.79MiB
243          webm       622x360    360p  459k , vp9, 30fps, video only, 96.26MiB
134          mp4        622x360    360p  616k , avc1.4d401e, 30fps, video only, 70.79MiB
17           3gp        176x144    small , mp4v.20.3, mp4a.40.2@ 24k
36           3gp        320x186    small , mp4v.20.3, mp4a.40.2
18           mp4        622x360    medium , avc1.42001E, mp4a.40.2@ 96k

43           webm       640x360    medium , vp8.0, vorbis@128k (best)



No comments:

Post a Comment