Course
Principles of Statistics for Data Analysis
Learn to describe and picture any dataset honestly, reason with chance, use the normal curve where it fits, and judge where data came from, all in a spreadsheet.
Before you begin
Recommended first: How Data Actually Works.
You work in a spreadsheet (Google Sheets or Microsoft Excel). If data is new to you, do the free How Data Actually Works first.
This is guidance, not a gate. You can enrol and start whenever you like.
What you will be able to do
- Say what kind of data each column holds (category, ordered category, count, measurement) and what that allows you to calculate.
- Summarise a dataset with the right measure of centre (mean, median, mode, weighted average) and spread (range, IQR, standard deviation, coefficient of variation), and say why you chose them.
- Describe where a value stands with percentiles and z-scores.
- Pick, make and read the right chart: dot plot, histogram, box plot, bar chart, line chart.
- Work out and explain chances from tables, including "not", "or", "and", "given", independence, base rates and expected values.
- Use the binomial and Poisson distributions to count chances, and the normal curve where it fits; check the fit with a Q-Q plot, and tame skew with logs.
- Tell a population from a sample, choose a fair sampling method, name the kinds of bias, and tell an observational study from an experiment.
Syllabus
Making sense of a pile of numbers
How spread out is it?
Same average, different story: the range
Paid
The middle half: quartiles and the IQR
Paid
Where do I stand? Percentiles and cumulative frequency
Paid
How far from the average? Variance in plain words
Paid
Standard deviation: the typical distance
Paid
A sample or the whole group? STDEV.S and STDEV.P
Paid
Comparing spread fairly: the coefficient of variation
Paid
Is this value unusual? Fences for outliers
Paid
Check yourself: how spread out is it
Paid
The shape of data, and the right picture
Data has a shape: the dot plot and the stem-and-leaf plot
Paid
Histograms: sorting values into bins
Paid
Same data, different bins
Paid
Lopsided data: skew
Paid
The box plot: five numbers in one picture
Paid
Comparing groups side by side
Paid
Pictures for categories
Paid
Change over time: line charts, percentage change and percentage points
Paid
Two numbers for each thing: a first look at scatter plots
Paid
Check yourself: the shape of data
Paid
Probability in plain language
What \"chance\" means
Paid
Chances from a table
Paid
Not, or, and
Paid
When one thing changes the chance: conditional probability
Paid
Chances in a row: the multiplication rule and tree diagrams
Paid
Independent or connected?
Paid
The base-rate surprise
Paid
Probability distributions and expected value
Paid
Check yourself: probability in plain language
Paid
Counting chances: the binomial and the Poisson
Counting successes: when the binomial fits
Paid
Exact chances with BINOM.DIST
Paid
What to expect from a count: mean and spread
Paid
Waiting for the first success: the geometric distribution
Paid
Counts per day: the Poisson distribution
Paid
Many tries make a bell
Paid
Check yourself: counting chances
Paid
The normal curve: where it fits and where it does not
From histogram to curve
Paid
The normal curve and its two dials
Paid
Most values sit close: the empirical rule
Paid
z-scores: how many steps from the middle?
Paid
Finding chances with NORM.DIST and NORM.INV
Paid
Is it normal? The Q-Q plot
Paid
When the bell does not fit
Paid
Taming skew with logs
Paid
Check yourself: the normal curve
Paid
Samples, bias and study design
The whole pot and the spoonful: population and sample
Paid
Fair spoons: random sampling
Paid
Other fair spoons: stratified, cluster, systematic and multistage samples
Paid
How bias sneaks in
Paid
Watching or testing? Observational studies and experiments
Paid
Making an experiment fair: control, random assignment, replication, blocking, placebo and blinding
Paid
Collecting data ethically: consent, privacy and anonymising
Paid
Capstone: the Harbourside water and health survey
Paid
Check yourself: samples, bias and study design
Paid
Final exam: Principles of Statistics
Paid