Can I trust this data?

What does this file look like, and is anything in it suspect before analysis?

The data. standard_data_profile_v2.csv, the published sample file for Data profile: 800 rows; columns account_id, monthly_spend, visits, days_open, region, plan, currency, last_note, and more. Sample data, not a customer's.

The calculation. Python, standard library only. This exact code ran against the file when this page was built, and printed the output below it.

import csv
rows = list(csv.DictReader(open("standard_data_profile_v2.csv")))
cols = list(rows[0])
ids = {r["account_id"] for r in rows}
print(len(rows), "rows,", len(cols), "columns,", len(ids), "distinct ids")
for c in cols:
    blank = sum(1 for r in rows if not r[c].strip())
    if blank:
        print(c, "is blank in", blank, "rows", f"({blank / len(rows):.0%})")
days = sorted(float(r["days_open"]) for r in rows)
print("days_open: median", days[len(days) // 2], "largest", days[-1])

800 rows, 9 columns, 800 distinct ids
last_note is blank in 230 rows (29%)
legacy_code is blank in 800 rows (100%)
days_open: median 400.0 largest 9000.0

What it says. There are as many distinct account ids as rows (800), so this is one row per account: a cross-section, not a panel. 1 column is empty in every row and can be dropped. The largest days_open is 9000 days, about 25 years, against a median of 400: a value worth checking before it pulls an average.

What it does not say. A blank is not always missing at random: if notes are left empty mostly for one kind of account, dropping those rows biases what follows. A profile flags such things; it does not decide them.

Run Data profile on this file or your own, or open its stored sample report.