How to Compute Key Statistics with NumPy and Pandas DataFrames

Using NumPy to Compute Statistics

The NumPy library provides functions for calculating a variety of statistical metrics. You can also use functions from scipy.stats for the same purpose. The table below lists common NumPy statistical functions.

Example: A school randomly selects 100 students and records their height and weight. Using the data, compute the height's mean, median, range, variance, and standard deviation, as well as the covariance and correlation between height and weight.

The computed covariance and correlation values are shown in the original source. The resulting statistics for height are: mean, median, range, variance, and standard deviation (values omitted here). The covariance and correlation between height and weight are also provided.

Using Pandas DataFrame to Compute Statistics

Pandas DataFrames offer a collection of built‑in statistical methods. The following list summarizes several useful functions.

count : Returns the number of non‑NaN entries.

mad : Computes the median absolute deviation.

mode : Returns the mode(s), i.e., the most frequently occurring value(s).

skew : Returns the skewness of the data.

kurt : Returns the kurtosis of the data.

quantile : Returns sample quantiles; by default it returns the 50th percentile (median).