Temperature Data Extraction (Open-Meteo API)

PROJECT TITLE:
Extract Historical Hourly Forecast Temperature Data (Open-Meteo)

PROJECT DESCRIPTION:
I require extraction of historical hourly forecast temperature data from the Open-Meteo Previous Runs API.

The goal is to retrieve archived hourly forecast temperatures for a fixed list of US locations from:

START DATE: 2024-01-01
END DATE: [INSERT TODAY’S DATE]

No UI, no dashboard, no hosting required.
This is a data extraction and structuring task only.

------------------------------------------------------------

DATA SOURCE:

Open-Meteo Previous Runs API
Base endpoint:
https://previous-runs-api.open-meteo.com/v1/forecast

------------------------------------------------------------

REQUIRED VARIABLES (HOURLY):

Retrieve the following hourly fields:

- temperature_2m
- temperature_2m_previous_day1
- temperature_2m_previous_day2
- temperature_2m_previous_day3
- temperature_2m_previous_day4
- temperature_2m_previous_day5

Timezone must be: UTC
Temperature unit must be: Fahrenheit

------------------------------------------------------------

LOCATIONS:

I will provide a list of 4 US locations including:
- Station name
- Latitude
- Longitude

All 4 must be included.

------------------------------------------------------------

DATE RANGE:

Start: 2024-01-01
End: [INSERT TODAY’S DATE]

Full continuous coverage required across entire range.

------------------------------------------------------------

REQUIRED OUTPUT STRUCTURE:

Deliver as CSV (UTF-8 encoded).

Each row must represent:
(station_id, timestamp_utc, lead_days)

Required columns:

- station_id (string)
- station_name (string)
- latitude (float)
- longitude (float)
- timestamp_utc (ISO 8601 format, UTC)
- lead_days (integer: 0–5)
- temperature_f (float)

Lead mapping:

lead_days = 0 → temperature_2m
lead_days = 1 → temperature_2m_previous_day1
lead_days = 2 → temperature_2m_previous_day2
lead_days = 3 → temperature_2m_previous_day3
lead_days = 4 → temperature_2m_previous_day4
lead_days = 5 → temperature_2m_previous_day5

------------------------------------------------------------

DATA INTEGRITY REQUIREMENTS (MANDATORY):

1. No duplicate rows.
Unique key must be: (station_id, timestamp_utc, lead_days)

2. No missing timestamps within API-available range.

3. All timestamps must be in UTC.

4. All lead_days layers (0–5) must be present for all timestamps returned.

5. Include an audit summary file showing:
- Total rows per station
- Total rows per lead_days layer
- Date coverage per station
- Count of null values (if any)

------------------------------------------------------------

EXTRACTION REQUIREMENTS:

- Data must be retrieved in safe date windows (e.g., 7–30 day blocks)
- Script must handle retries and timeouts
- Full coverage must be verified before delivery

------------------------------------------------------------

DELIVERABLES:

1. Final clean CSV dataset
2. Audit summary file
3. Python extraction script used
4. Short README explaining extraction method

------------------------------------------------------------

ACCEPTANCE CRITERIA:

Work will be accepted if:

- All 4 stations included
- Full date range covered
- No duplicate composite keys
- Lead layers 0–5 present
- Audit totals match dataset

This is a data extraction task only. No analysis required.