README improvements

[waldyrious]

• Add a simpler introductory example (inspired by this Discourse post by @Satvik)
• Rename "Example" section to "Usage" and make the two next sections into subsections of it
• Normalize headings to sentence case
• Reword a couple sentences to adjust to the package rename (from PipelessPipes.jl to Chain.jl)
• Add a link to DataFrames.jl

[jkrumbiegel]

In principle I'm not opposed to adding a simple non-DataFrame example. But _ .+ 1 and _ .^ 2 show the same syntax twice, which doesn't show off the range of options.

This for example shows a couple different things you can do:

@chain 1:10 begin
	  _ .^ 2
	  mod.(5)
	  filter(iseven, _)
	  sum
end

Maybe there is a more didactic example than this.

[jkrumbiegel]

What about this:

julia> names = ["ursula", "alice", "iris", "olga", "erica"]
5-element Vector{String}:
 "ursula"
 "alice"
 "iris"
 "olga"
 "erica"

julia> @chain names begin
    filter(contains("a"), _)
    @. uppercasefirst
    sort
    join(" and ")
    "$_ have an 'A' in their name."
end
"Alice and Erica and Olga and Ursula have an 'A' in their name."

[waldyrious]

That's a good point — but I believe it makes sense for the very first example to aim for a demonstration of the basic structure of a chained block, rather than showcasing the variety of possibilities. With that in mind, I would maybe start with the first example you offered above, and show the second one next. This way we can keep the advantages of both. WDYT?