Ask two questions about a class of 30 students: how many have a pet, and how tall each student is. The first answer is a count, 12 or 13, never 12.4. The second is a measurement: a height might be 158 cm, or 158.3 cm, or 158.27 cm with a finer tape, and any value in between is possible. Counts of people, objects or events are discrete. Measurements of time, length, mass and temperature are continuous.
The test is whether a value strictly between two possible values is also possible. Between 7 cars and 8 cars there is no possible count. Between 20.1°C and 20.2°C sits 20.15°C, and between those sit more, because air warms smoothly through every value. Graphs follow the same rule. Tickets sold each day are plotted as separate dots, while temperature through the day is drawn as a connected curve. Joining the ticket dots would suggest values at 'day 2.5' that do not exist.
Recording can disguise a continuous quantity as discrete. A form asks for age in whole years, a thermometer displays 21°C, a stopwatch shows 12.4 s. The numbers you see are rounded, but the age, temperature and time underneath passed through every value in between. The opposite trap exists too: money looks smooth, but an account balance changes in whole cents, so strictly it is discrete with a step of $0.01. Classify by which values are possible, not by what was written down.
Think of it like this. Discrete is a staircase and continuous is a ramp. On stairs you can stand only on whole steps; on a ramp you can stop at any height, and between any two heights there is another spot. A staircase with tiny steps, like cents in money, is often treated as a ramp because the steps are too small to matter.
Core idea. Discrete data lands on separate values; continuous data can take any value in a range.