Post-stratified results for a two-wave customer survey with repeat respondents and mixed income units

# Weighted results for a two-wave customer survey Inputs: `responses.csv`, `waves.json` (per wave: the unit of the income column), `population.json` (each region's share of the customer base). Rules: - One person may have answered more than once. Emails are case-insensitive and may carry stray spaces. Keep only each person's latest submission by actual time; submitted_at carries different UTC offsets. - Drop respondents (after de-duplication) whose age is blank or under 18. - In q1-q3, the code 99 means "don't know": leave it out of that question only. - Income is in the unit given for the response's wave in waves.json: "USD" or "kUSD" (thousands of USD). Report income in USD. - Weight for every kept respondent in region g = population share of g / (kept respondents in g / all kept respondents). - Weighted mean = sum(weight x value) / sum(weight) over the kept respondents who have a valid value. Round q means half-up to 3 decimals and income half-up to 2 decimals. Write `out/survey.json`: `{"n": <kept respondents>, "region_n": {"<region>": <int>, ...}, "weighted_means": {"q1": <n>, "q2": <n>, "q3": <n>}, "weighted_mean_income_usd": <n>}`. q means are checked to within 0.0005. ## How this run works - Work on your own: nobody will answer questions in this task, and writing to `ask/` counts against the run. - Write only under `out/`. Create `out/` if it is missing. Inputs are synthetic; don't fetch anything from the network. - Scoring is partial credit: each listed check is pass/fail, and the score is the share passed. Numbers are compared to within 0.005 unless the task says otherwise.