Organisation of this repo (examples, benchmarks, etc)

The current repo organisation is very messy and needs better re-organisation. Below are notes

All examples should be implemented in both Turing.jl and Stan.jl
- Each of them needs a file for modelling and a file for running
Currently each model has its own data file, which may not be ideal as
- We probably want to share data between models when the example base gets larger
- Reusing data according to model names seems weird and unclear.
Any synthesised data should has its corresponding simulation file

Given that we have LDA and MoC examples, the repo organisation (benchmark related things excluded) below is proposed:

/models
  /lda
    /turing.model.jl
    /turing.run.jl
    /stan.model.jl
    /stan.run.jl
  /moc
    /turing.model.jl
    /turing.run.jl
    /stan.model.jl
    /stan.run.jl
/data
  /topic.data.jl # this contains helper functions to load .jld file into Julia;
                 # also provides an consistent way to use data which is provided in Julia directly
  /topic.sim.jl
  /topic.jld

TuringLang / papers

Organisation of this repo (examples, benchmarks, etc) #4