Data sets are a key part of a modern society and the ability to store them in an open, accessible format makes them easy to manipulate.
However, a new paper by researchers at the University of Michigan and Georgia Tech sheds light on how data sets are used by different groups.
The researchers’ study, published in the Proceedings of the National Academy of Sciences, analyzed the data for the United States in 2014 and 2015.
The study looked at a total of nearly 12 million datasets, of which 4 million were publicly available.
They analyzed which data types were most commonly used by people, what kinds of data were being used, and how they were structured.
It also looked at the impact of public data on privacy and security.
The report found that data sets tend to be less public than they used to be.
For example, in 2014, almost 1.5 billion public datasets were available.
But this number decreased to 1.1 billion in 2015.
That decline is mainly due to the growth of data-mining services such as Google Cloud, Facebook, and Twitter.
The trend for publicly available data has also been changing.
Data sets that are publicly available now often are less secure than they once were.
In 2014, the majority of the data was stored on public servers.
The new paper, however, finds that data that is used on public-facing sites is generally less secure and less secure as a result.
The paper notes that in 2014 public data were used for things like reporting and statistics, but those uses have decreased in recent years.
This may be because companies are moving away from these kinds of services, and as a consequence there is less demand for public data.
More recently, some governments have been using the data to try to track criminals.
This has led to concerns about data security, and in 2016, the European Commission proposed new rules to protect data that has been used in a way that is deemed to violate EU law.
The research has been funded by the National Science Foundation.
The article was written by Emily Kelleher and is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.