Update README.md

data/README.md

@@ -1,9 +1,20 @@
 # Data
 
+## Overview
 This directory contains two important subdirectories: `raw/` and `processed/`. `raw/` will house raw data files that will be fetched when running `make data` from the top level project directory. `raw/` should expect to be filled with approximately 37.5 GBs of data. `processed/` will house files processed and generated by native project scripts. `processed/` should expect to be filled with approximately 190 GBs of data.
 
 If you want the data to instead be stored on another drive, delete `raw` and `processed` with `mkdir raw && mkdir processed` and instead create symbolic links with those names to the path you want your data to be located at. For example, if you want your raw data to be located at `/Volumes/Passport/raw`, then you can create a symbolic link via:
 
     $ ln -s /Volumes/Passport/raw raw
 
 and analogously for `processed`
+
+## Data file meaning
+There are several different types of data files. We outline the naming convention below.
+- `raw/subi_runj_raw.nii.gz` is the compressed raw data file for subject **\#i** run **\#j**
+- `raw/subi_runj_raw.nii` is the uncompressed raw data file for subject **\#i** run **\#j**
+- `processed/subi_rcds.npy` contains concatenated raw data across all 8 runs for subject **\#i**
+- `processed/subi_rcds_2d.npy` contains concatenated raw data across all 8 runs for subject **\#i** reshapes to two dimensions (voxels by time)
+- `processed/subi_rcds_smoothed_j_mm_2d.npy` contains gaussian smoothed data for subject **\#i** smoothing with FWHM of **\#j**
+- `processed/subi_rcds_smoothed_j_mm_2d.npy` contains gaussian smoothed data for subject **\#i** smoothing with FWHM of **\#j** reshaped to two dimensions (voxels by time)
+- `processed/subi_subj_correlation.npy` contains pearson correlation between subject **\#i** and subject **\#j**