11. People from different places and different
backgrounds tend to produce different sorts of
information. And so we risk ignoring a lot of
important nuance if relying on big data as a
social/economic/political mirror.
Mark Graham, “Big data and the end
of theory” The Guardian. March 2012
Just in time manufacturing – kanban. 1950’s Toyota looked at how supermarkets handle inventory delivering produce just at the time customers in the store needed. In manufacturing large inventories were kept in storage and analysis was done on what volumes levels should be held based on historical data. In contexts where supply time is lengthy and demand is difficult to forecast, the best case can be to respond quickly to observed demand.Kanban changed this with dynamic job queues and just in time delivery. Can the information economy get the same gains the manufacturing economy got with just in time methods.
The heijunka box is generally a wall schedule which is divided into a grid of boxes or a set of 'pigeon-holes'/rectangular receptacles. Each column of boxes representing a specific period of time, lines are drawn down the schedule/grid to visually break the schedule into columns of individual shifts or days or weeks. Coloured cards representing individual jobs (referred to as kanban cards) are placed on the heijunka box to provide a visual representation of the upcoming production runs.The heijunka box makes it easy to see what type of jobs are queued for production and for when they are scheduled. Workers on the process remove the kanban cards for the current period from the box in order to know what to do. These cards will be passed to another section when they process the related job.
Today’s data warehouses
Big data originated with batch operations with massive amount of data. In effect performing valuable post mortems on our digital exhaust
Increasing the data we create is coming in real time
It is not just social media but sensors, mobile, all globally instrumented
Real time analytic tools are quickly emerging but are out old analytical methods appropriate
In 1749 the German scholar Gottfried Achenwall suggested that since this ‘science’ [the study of society by counting] dealt with the natural ‘states” of society, it should be called Statistik.
Two trends using big data as a reflection of society and data is a living organism and a post mortem is for things that are dead.
The downside to reductionist approaches. Is the end of theory premature?
Any event with a very low probabilitythat occurs gives us a great deal of information, whereas when an event with a highprobability occurs, this is less of a surprise and gives us correspondingly less information but greater certainty. Information thus varies inversely with probability.You should call it entropy, for two reasons. In the first place your uncertainty function has been used in statistical mechanics under that name, so it already has a name. In the second place, and more important, no one really knows what entropy really is, so in a debate you will always have the advantage!
Entropy measures of Vogue covers based on the complexity of the pixel variations in the graphic layout
This function has many attractive properties for describing spatial distributions.Here, we initially assume that the probability piis proportional to some count ordensity of spatial activity, such as population in a zone i that might be a censustract. If all the population were located in a ‘‘mile-high building’’ such as the oneproposed for a town of 100,000 people in 1956 by Frank Lloyd Wright (Rybczynski2010), then pi 5 1 and pk ¼ 0; 8k ¼6 i, and the entropy would be at a minimum,with Hmin 5 0. If the population were evenly spread throughout the tracts as pi ¼ 1=n; 8i, then the entropy would be at a maximum, with Hmax ¼ log n. (Batty 2011)
Max entropy over space and time
Traditional statistical approaches to dynamic data. Categorical splits.