
228 • Big Data Analytics
e reality is statistics are oen poorly used. We suggest some guide-
lines to lead the reader to spot poor usage. When managing analysts or
contracting with a rm showing o the results it has attained with data, it
is worth asking questions such as
• Are we basing our analysis on the mean or median? (And remember
that the mean can lose a great deal of its meaning if the distribution
is not symmetric!)
• Is our distribution based on individual data points or collections of
points (think of the example with dice)?
• What distribution did we use?
• What is the dispersion of the data?
• If there is a threshold past which ...