Data Summaries Flashcards
types of data, numerical summaries, graphical summaries
What are the two categories of variables?
- quantative (numerical value, continuous or discrete)
- qualitative (non-numerical value, ordinal (relative values) or nomial (unordered, distinct by name only))
What is the basic stucture of a statistical model?
outcome = model + error
outcome - signal + noise
What is the difference between the sample mean (μ^) and the population mean (μ)?
Sample mean is the mean of a sample (/n) while population mean is for the whole population (/N).
What is the interquartile range (IQR)?
75th percentile - 25th percentile
How is a standard deviation done on a calculator (Casio)?
setup > stat > AC > shift + 1 > data > enter data > AC > shift + 1 > var >
What is a histogram?
Data partitioned into distinct bins.
How is skewness worked out, and what types does it have?
Diagnosed visually.
- right skewed: long right tail (mean > median)
- left skewed: long left tail (mean < median)
- symmetrical (median = mean)
- bi-modal (two peaks)
What is a box plot?
- box spans IQR
- median is line across the box
- “whiskers” extend to 1.5x the IQR beyond the box
- values beyond whiskers are outliers
- notched box plots (notches represent plausible values for median)
What is a violin plot?
- combine box plots with smooth, sideways histogram
- displays the median as a dot
What is a quilt plot?
- useful to summarise geo-referenced data
- partition x- and y-coords into sections
- average value represented in each box as a colour
What are bar plots and pie charts useful for?
Easy and clear way to illustrate frequency info across different groups
What are some important things to include when constructing plots?
- clearly label axes, units and title, provide key (if needed)
- use suitable scale, plot the origin
- consider if a table is more suitable
What are some important things to include when constructing tables?
- make it clear and simple, main numbers for comparrison close to each other
- arrange rows and columns in natural order
- choose convinient units, include title and brief explanation of data shown
- consider splitting complicated tables into two smaller tables
- round to appropriate sig figs (~2)