13 videos · 78 minutes · every word written out
Every video, with the answer written out
One idea per video, in plain English, no prior maths assumed. Each page below opens with the short answer to the question in the title, then the chapters, then the full transcript — so you can read it instead of watching if that suits you better. Want them in teaching order rather than newest first? That is the free course.
-
Video 13 · 5:26 · Oct 2026
Misleading Graphs: Why the Axis Matters
How can a graph mislead you without changing the numbers?
Through the axis. In a bar chart the length of the bar is the number, so the axis has to start at zero. Start it higher, and a small difference looks huge: in the example, bus 4 is 6.5 minutes late on average and bus 7 is 6.0, only half a minute or about 8% more, yet on an axis starting at 5.5 minutes bus 4's bar is twice as tall. Zoom out too far and a real difference can look like nothing. Line charts may start above zero, because they show change, but your eye still makes the change look bigger, and two scales on one chart can make almost any two lines look like partners. In experiments, more than four in five people saw a bigger difference on a cut-off axis, even right after it was explained to them.
-
Video 12 · 7:02 · Sep 2026
Correlation vs Causation: Do Big Feet Read Better?
Why does correlation not imply causation?
Two things can move together without one causing the other. Every link comes with three stories: A causes B, B causes A, or a third factor pushes on both. In the example, children with bigger feet read better (r = +0.75), but only because older children have bigger feet and read better; inside each age group the correlation drops to about zero. A correlation alone cannot show a cause. The most direct way is an experiment that hands out the possible cause by lottery, so every third factor ends up in both groups about equally.
-
Video 11 · 6:46 · Sep 2026
Correlation: What r Really Tells You
What does the correlation coefficient r actually measure?
r measures how tightly the dots in a scatterplot hug a sloping straight line. It runs from minus one to plus one: the sign gives the direction and the size gives the strength, so minus 0.8 is just as strong as plus 0.8. A value near zero means there is no straight-line trend, not that there is no link: a perfect curve, like the height of a thrown ball over time, can have r = 0. Always look at the cloud as well, because your eye misjudges r and a single far-out dot can move it a long way.
-
Video 10 · 6:58 · Sep 2026
Scatterplots: Two Variables at Once
How do you read a scatterplot?
Each dot is one case with two measurements, so the pairing is the whole point of the picture. Read it in four steps: the direction (do the dots climb together, trade places, or go nowhere), the form (a straight band or a bend), how tightly the dots keep to that band compared with how far the band travels, and the dots that do not fit. A scatterplot shows a pattern, never a reason: it cannot tell you which of the two numbers came first.
-
Video 9 · 5:43 · Sep 2026
Variance: Why We Divide by n − 1
What is the difference between variance and standard deviation?
They are the same measurement in two units. The variance is the average squared distance from the mean, so it comes out in squared units; the standard deviation is its square root, which brings you back to the unit you started in. You divide by one less than you counted because a sample measured from its own average always understates how spread out it really is.
-
Video 8 · 6:21 · Sep 2026
Histograms: Bin Width Changes the Story
What is a histogram, and how do you read one?
A histogram cuts a list of numbers into bins of equal width and stacks how many land in each. The height means one thing: the count. Unlike a bar chart the bars touch, because the axis underneath is a continuous scale. The catch is that the bin width is a choice, and different widths can tell different stories about the same numbers.
-
Video 7 · 6:51 · Sep 2026
The Log Transformation: One Axis Trick
What is a log scale, and when do you use one?
A log transformation changes the ruler, not the data. On a logarithmic axis equal steps mean equal ratios — ×10 takes the same space every time — so values squashed against one wall open up into something readable, often a bell. It is the standard fix for data spanning several orders of magnitude, such as income or population.
-
Video 6 · 6:22 · Sep 2026
Skewed Distributions Explained Simply
What does it mean when data is skewed?
A distribution is skewed when it has a hill on one side and a long tail on the other — and the name follows the tail, so right-skewed means the tail stretches to the right. In right-skewed data the mean sits above the median, because the tail drags the balance point with it. Income, waiting times and house prices nearly all lean this way.
-
Video 5 · 5:20 · Aug 2026
Mean vs Median: Which One to Use When?
Mean or median — which one should you use?
The mean is the balance point: add everything up and divide by how many. The median is the middle person: sort the list and take the one in the centre. On symmetric data they agree. When a few extreme values pull on one end the mean follows them and the median hardly moves, which is why income is reported as a median.
-
Video 4 · 5:01 · Aug 2026
The Boxplot Explained Simply
How do you read a boxplot?
A boxplot folds five numbers into one small picture: minimum, first quartile, median, third quartile and maximum. The box covers the middle half of the data, the line inside it is the median, and the whiskers reach to the last value within 1.5 × IQR. Anything beyond that is drawn as its own dot.
-
Video 3 · 4:50 · Aug 2026
Percentiles & IQR Explained Simply
What is a percentile, and what is the IQR?
A percentile says what share of the data falls below a value: the 90th percentile is the point 90% of the group sits under. You get it by sorting and counting, so it assumes nothing about the shape of your data. The quartiles Q1, median and Q3 cut the sorted list into four, and the IQR — Q3 minus Q1 — is the width of the middle half.
-
Video 2 · 4:53 · Aug 2026
Z-Scores Explained Simply: How Rare Is 2 Metres?
What is a z-score?
A z-score says how many standard deviations a value sits from the mean, which turns every bell curve into the same ruler. A z of +2 means two standard deviations above average, whatever you are measuring. That is how you compare a height with an exam score, and how you say how rare something really is.
-
Video 1 · 6:07 · Aug 2026
The Normal Distribution, Explained Simply
What is a normal distribution?
A normal distribution is the bell-shaped pattern you get when many small, independent influences add up: heights, measurement errors, reaction times. Two numbers describe it completely — the mean μ (mu), which sits at the peak, and the standard deviation σ (sigma), which sets how wide the bell is. Know both and you know where almost everyone falls.
Up next: Types of Data: Categorical vs Continuous — What is the difference between categorical and continuous data?