Www.julialang.org: DataFrames.jl 1.0 blog post

[bkamins]

Following the request by @quinnj (following @ViralBShah) I open the issue here to start discussing what to put on the blog post about DataFrames.jl.

Here is a quick list of things that come to my mind that could be shown:

1. Why it took so long to reach 1.0 release: history of DataFrames.jl and its consequences (many rounds of deprecation cycles)
2. Composability made possible by Tables.jl and DataAPI.jl (examples: CSV.jl & Arrow.jl, PrettyTables.jl, PooledArrays.jl & CategoricalArrays.jl, StatsPlots.jl)
3. Why it is hard to make DataFrames.jl fast: ways to handle type instability and compilation latency (e.g. essentially every combine call in normal user's workflow has to be compiled because of anonymous functions)
4. Different coding styles offered: imperative, declarative (in-place operations and Chain.jl examples)
5. Integration with Julia Base: indexing, views
6. Design behind GroupedDataFrame
7. Some performance benchmarks
8. What does it mean for the users that we have 1.0 release and what is to be done yet post 1.0 release

However, I think we should discuss what we want to cover in the post (as the list here is a bit haphazard)

[ViralBShah]

It would be nice to have some discussion on the philosophy behind the package and its design.