A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Descriptive statistics summarizes observed data; probability provides a mathematical framework for reasoning about uncertainty and future observations. The bridge between them is relative frequency: when you compute that 30% of data values fall in a certain range, you are implicitly estimating the probability that a new observation will land there. Formalizing this connection means moving from "here is what happened" to "here is what we expect to happen." This transition is the conceptual gateway to random variables, expected value, and all of inferential statistics — where observed data is used to draw conclusions about the underlying process that generated it.
Start with a frequency distribution from real data and convert frequencies to relative frequencies. Discuss how those relative frequencies behave like probabilities. Then formally introduce the probability axioms and show that relative frequency distributions satisfy them. Use simulation (rolling dice many times, sampling from a dataset) to illustrate the convergence of relative frequency to theoretical probability.
Descriptive statistics is the art of summarizing what happened. A histogram, a mean, a standard deviation — these are tools for compressing observed data into digestible form. But they say nothing about what will happen next. To make predictions and draw general conclusions, you need a different kind of machinery: probability theory, which you already know through the probability axioms. This topic is about recognizing that the two fields are not independent subjects — they are connected by a single bridge called relative frequency.
Here is the bridge in action: suppose you record the outcome of rolling a fair die 600 times. Roughly 100 of those rolls land on a 3 — a relative frequency of about 100/600 ≈ 0.167. Compare that to the theoretical probability of rolling a 3: 1/6 ≈ 0.167. The relative frequency from the data approximates the probability from the model. As the number of observations grows, the approximation improves — this is the Law of Large Numbers, which you will prove rigorously later. The observation that relative frequencies converge to probabilities is what connects your descriptive summaries to the probability axioms you already know.
This connection runs in both directions. Statistics to probability: when you see a relative frequency in data, you can use it as an estimate of an underlying probability. A health researcher who observes that 23% of patients in a study respond to a treatment estimates the probability of response at 0.23. Probability to statistics: when you know the probability model, you can predict what data should look like. If a coin is fair (probability 0.5 per flip), you expect roughly half the observations to be heads. These two directions define the two disciplines: statistics reasons from observed data to an unknown model (backward); probability reasons from a known model to expected observations (forward).
The key conceptual shift here is from data as a collection of fixed facts to data as a sample from a broader process. Descriptive statistics treats your dataset as the entire story. Probabilistic thinking treats your dataset as one realization of a random process — one possible roll of the dice. Once you make this shift, questions like "how confident am I in this estimate?" and "how different could this result have been?" become meaningful. That is the gateway to all of inferential statistics: you use observed data to say something about the underlying probability distribution that generated it.