Herman Hollerith filed his first patent application on September 23rd, 1884. It was entitled "Art of Compiling Statistics," and was a description of how a machine could mechanically process at speed punched cards containing information. The idea was simple, but revolutionary. Data could be input by making or not making a hole in the card. Thus in one row, a hole might indicate married, whereas the absence of one might mean unmarried. In the next row a hole might mean North of England, while the absence of on might mean South of England. The cards, fed mechanically at speed, would encounter teeth that slotted them into appropriate columns. A pile of cards in one tray might contain all those registered as married, living in North of England.
The data could be registered on cards with electro-sensitive pens, to be sorted electromechanically. Hollerith gained his doctorate at Columbia University with his "Electric Tabulating System," and went on to found The Hollerith Electric Tabulating System which took the world into semiautomatic data processing. His Tabulating Machine Company merged with others to become the Computing-Tabulating-Recording Company, renamed IBM in 1924.
His system proved ideal for the US Census Bureau, and enabled them to process date two years faster than the previous census had taken. Census bureaux around the world leapt in to buy his machines, along with insurance companies and others that lived and traded on bulk data. Hollerith's machines were used for censuses in England, Italy, Germany, Russia, Austria, Canada, France, Norway, Puerto Rico, Cuba, and the Philippines, as well as in the US.
When my uncle was chief wages clerk at a UK steelworks, later wages manager, Hollerith punch card machines were still in use in the early 1960s, and I sometimes helped him enter data on the cards for subsequent mechanical sorting. Electromechanical sorting was eventually superseded by electronic sorting, and we are now in a world where computers do in seconds what took sorters and scribes years to do previously.
Big data remains a controversial field. In insurance it can enable companies to select out or charge high premiums to those whose data indicates them to be bad risks. Political campaigns can direct their messages to individuals based on what their data indicates will elicit a favourable response. Obviously, privacy is an issue, but the advantages of being able to process data rapidly seem much greater than any downside. Going through data at speed can help us to predict all kinds of unpleasant stuff that would previously have seemed to come out of the blue, and in predicting it, can perhaps enable us to avoid the unpleasant things we might otherwise have fallen victim to.
It seems unlikely that we'll ever be into the "psychohistory" of Isaac Asimov's Foundation, where algorithms based on vast numbers will enable us to predict specific future events in detail. This is unlikely, largely because human beings, unlike atoms and stars and rolling bodies, have minds of their own, and can change them. It is possible, however, that broad patterns might be discerned from the data. Humans are wonderfully creative beings, fortunately, and will no doubt find ways of using to advantage the data revolution that Herman Hollerith set in motion with his 1884 punch card machine.