Then there’s “other” data: 500 hours of video us uploaded to YouTube per minute.
No organization can go around watching all of this video data. Which parts are newsworthy?
Content could surface much later or could surface through later research.
Hornbjargsviti lighthouse recorded the weather every three hours for years creating lots of data.
And that was just one of hundreds of sites that recorded this type of data in Iceland.
Lots of this data is lost. Much that has been found was by coincidence. It was never thought to archive it.
This type of weather data could be very valuable to researchers later on.
There was also a large archive of Icelandic data that was found.
Showing a timelapse of Icelandic earthquakes https://vimeo.com/24442762
You can watch the magma working it’s way through the ground before it makes it’s way up through the land.
National Geographic featured this video in a documentary.
Sometimes context is important when it comes to data. What is archived today may be more important later.
As the economic crisis unfolded in Greece, it turned out the data that was used to allow them into EU was wrong.
The data was published at the time of the crisis, but there was no record of what the data looked like 5 years earlier.
Only way to recreate the data was to take prior printed sources. This is usu only done in extraordinary cirucumstances.
We captured 150k+ data sets with more than 8 billion “facts” which was just a tiny fraction of what exists.
How can we delve deeper into large data sets, all with different configurations and proprietary systems.
“There’s a story in every piece of data.”
Once a year energy consumption seems to dip because February has fewer days than other months. Plotting it matters.
Year over year comparisons can be difficult because of things like 3 day weekends which shift over time.
Here’s a graph of the population of Iceland. We’ve had our fair share of diseases and volcanic eruptions.
To compare, here’s a graph of the population of sheep. They outnumber us by an order(s) of magnitude.
In the 1780’s there was an event that killed off lots of sheep, so people had the upper hand.
Do we learn more from reading today’s “newspaper” or one from 30, 50, or 100 years ago?
There was a letter to the editor about an eruption and people had to move into the city.
letter: “We can’t have all these people come here, we need to build for our own people first.”
This isn’t too different from our problems today with respect to Syria. In that case, the people actually lived closer.
In the born-digital age, what will the experience look like trying to capture today 40 years hence?
Will it even be possible?
Machine data connections will outnumber “people” data connections by a factor of 10 or more very quickly.
With data, we need to analyze, store, and discard data. How do we decide in a spit-second what to keep & discard?
We’re back to the father-in-law and mother-in-law question: What to get rid of and what to save?
Computing is continually beating human tasks: chess, Go, driving a car. They build on lots more experience based on data
Whoever has the most data on driving cars and landscape will be the ultimate winner in that particular space.
Data is valuable, sometimes we just don’t know which yet.
Hoarding is not a strategy.
You can only guess at what will be important.
“Commercial use in Doubt” The third sub-headline in a newspaper about an early test of television.
There’s more to it than just the web.