Saw this:
https://informedinsport.com/new-blog/special-post-the-illogic-of-being-data-driven
— Nate Silver
So, does data have a way to speak for itself, is all interpretation and meaning imbued solely by the people analyzing the data?
The problem with this question is that it misses a level of analysis.
True, meaning is only conferred through humans, outside the system, mapping external systems to the data. Semiotic definition of meaning holds.
However, since there are multiple potential interpretations we need a step to maximize the Information in the data before communicating its message to the people.
True it requires that people impose a metric, a type of a prior assumption, yet it is not an assumption on interpretation rather an assumption of the field of analysis. My favorite metric is that coincidence is informational. Thats it. Hence two observations that co-occur are related to each other. The basic assumption is that 'time' is a shared metric across all observations.
Then the data does speak for itself, it communicates a highly informational message. We still need to hear the message and interpret it to imbue the message with meaning. However, since the data has spoken well, with high informational content, the message will be clear.
Should I give an example?
https://informedinsport.com/new-blog/special-post-the-illogic-of-being-data-driven
The numbers have no way of speaking for themselves. We speak for them. We imbue them with meaning
So, does data have a way to speak for itself, is all interpretation and meaning imbued solely by the people analyzing the data?
The problem with this question is that it misses a level of analysis.
True, meaning is only conferred through humans, outside the system, mapping external systems to the data. Semiotic definition of meaning holds.
However, since there are multiple potential interpretations we need a step to maximize the Information in the data before communicating its message to the people.
True it requires that people impose a metric, a type of a prior assumption, yet it is not an assumption on interpretation rather an assumption of the field of analysis. My favorite metric is that coincidence is informational. Thats it. Hence two observations that co-occur are related to each other. The basic assumption is that 'time' is a shared metric across all observations.
Then the data does speak for itself, it communicates a highly informational message. We still need to hear the message and interpret it to imbue the message with meaning. However, since the data has spoken well, with high informational content, the message will be clear.
Should I give an example?
Comments
Post a Comment