# astroML chapters 1--4 notebooks

## Generic comments

- Link often to "upstream" resources. Whether it is stats resources or programming or tutorials. Maybe add a section at the bottom/top of all the notebooks?
- Some packages have steeper learning curves than others. Definitely link to their "getting started" tutorials, e.g. pymc3, but also we should link to scikit-learn's MOOC and the numpy tutorials, Jake's book, rather than try to cover all that material on our own.

### chapter1 -- Introduction

##### datasets
- Datasets are described when used in later notebooks, if repetition is preferred, a notebook going through all the datasets

##### Astronomy visualization
- A notebook that introduce the typical astronomical plots, that we use a lot later in the book. E.g. colour-magnitude, colour-colour diagrams, spectrum, light curves (both using time and phase), colour-redshift, healpix projection, etc. These are not necessarily covered in the text itself.

### chapter2 -- Fast Computation and Massive Datasets

- the figures are illustrations here, but all the code examples from the text can be worked into one notebook.

### chapter3 -- Probability and Statistical Distributions

##### Random variables
##### Common Univariate distributions
- would be nice to have a notebook for each distribution for easy referencing, but those would be very short. Maybe divide up the 12 into 2-3 notebooks? All distributions should have the same detailedness, even if it means repetition. Definition, a plot, and maybe an example or two mentioning where they are used. At this point we may also want to cross link to other resources, either educational, but definitely to code implementations to e.g. numpy and scipy stats, and pymc3. (We can link to more, but I would stay with packages that we also use.)

##### Central Limit Theorem
This may go into the random variables notebook, depending how long that would be
##### Bivariate and multivariate distributions
This includes 3.6 -- Correlation Coefficients, too.
##### Arbitrary distributions

### chapter4 -- Classical statistical inference

##### Maximum likelihood estimation, confidence intervals. We don't have illustrations

##### Goodness of Fit and model comparison

##### Expectation maximization
- Gaussian mixture models have been covered in chapter6 already

##### Confidence estimates

##### Comparison of distributions

##### Histograms
- Material for this comes from other chapters, too, e.g. ch5

### chapter5 -- Bayesian statistical inference

(we skipped these hoping to reuse something built for notebooks and really good from Gwen or Daniela)

##### Introduction to Bayesian

##### Model selection, information criteria