Skill: dikw-explore

Profile raw data, assess quality, note prior work. First phase of any DIKW session.

On invocation, use $ARGUMENTS to get the project path. If not provided, look for data files in the current working directory under source/raw/.

Steps

Find the project directory and data files:
- If $ARGUMENTS contains a path, use it as project_dir
- Otherwise use cwd as project_dir
- List all files in {project_dir}/source/raw/
Profile every data file:
- Use python3 with pandas
- For each file: print shape, columns, dtypes, null counts, sample rows
- Check for duplicates, outliers, missing values
Check for prior analysis:
- Look in {project_dir}/reports/ for existing D/I/K/W reports
- Note what levels already have work
Write exploration notes:

Profile raw data, assess quality, note prior work. First phase of any DIKW session.

On invocation, use $ARGUMENTS to get the project path. If not provided, look for data files in the current working directory under source/raw/.

Find the project directory and data files:
- If $ARGUMENTS contains a path, use it as project_dir
- Otherwise use cwd as project_dir
- List all files in {project_dir}/source/raw/
Profile every data file:
- Use python3 with pandas
- For each file: print shape, columns, dtypes, null counts, sample rows
- Check for duplicates, outliers, missing values
Check for prior analysis:
- Look in {project_dir}/reports/ for existing D/I/K/W reports
- Note what levels already have work
Write exploration notes: