data-explorer

name	data-explorer
description	Reads a dataset from data/clean/, describes its structure, suggests summary statistics, and flags data quality issues. Student must provide the file path in the same message.

Data Explorer Skill

Use this skill when the student wants to understand a dataset or identify data quality issues.

Input required

The student must provide the file path (relative to the project root, within data/clean/) in the same message as /data-explorer.

Example: /data-explorer data/clean/firm_panel.csv

If no path is provided, ask for it before proceeding.

Rules

• Read docs/student-profile.md to understand the research question and methodology.
• Read the file using the Read tool.
• For CSV files: parse the first row as headers and analyse structure from the first ~100 rows visible in the Read output.
• Do not modify the file.
• Suggest summary statistics the student should include in thesis/04_data.md.
• Generate Python code the student can run to produce the suggested outputs.

Output format

Data Explorer: [filename]

Schema

Column	Inferred Type	Non-null Count	Notes
[column name]	[int/float/str/date]	[count if visible]	[any observation]

Data Quality Flags

List any of the following if detected:

• Duplicates: [description]
• Missing values: [which columns, approximate frequency]
• Outliers: [which columns, describe]
• Implausible values: [e.g. negative prices, future dates]
• Date range: [if date column present]

Suggested Summary Statistics

Given the research question ([quote from student-profile.md]), the student should report:

• Variable: mean, standard deviation, min, max, N

Suggested Python Code

Paste this into code/04_tables_figures.py and adjust as needed:

import pandas as pd
from pathlib import Path

df = pd.read_csv("data/clean/[filename]")

# Summary statistics
summary = df[["var1", "var2"]].describe().T[["count", "mean", "std", "min", "max"]]
print(summary.round(3).to_string())

# Missing values
print(df.isnull().sum()[df.isnull().sum() > 0])

# Duplicates
print(f"Duplicate rows: {df.duplicated().sum()}")

Connection to Thesis

[1–2 sentences on how this dataset connects to the stated research question, and what the student should highlight in thesis/04_data.md.]