climate-indepth-analysis 0.1.0

A growing toolkit for climate data quality checks, EDA, and analysis.

Climate Indepth Analysis

climate-indepth-analysis is a Python package for climate data quality checks, exploratory data analysis, and future climate analysis workflows.

This first release focuses on daily station climate data EDA. It checks how clean a station dataset is by creating a full continuous daily calendar for every station before calculating missing values. This is important because many climate station files skip entire dates instead of storing those dates with NaN values.

Current features

• Count the number of climate stations in a file
• Support custom station ID columns such as STATION, station, ID, station_id, or site_no
• Support custom date columns such as DATE, date, datetime, or time
• Use default climate variables PRCP, TMAX, and TMIN
• Allow users to choose other variables
• Create a full station-date calendar before missing-data calculation
• Detect missing dates that are absent from the raw file
• Summarize missingness overall, by station, and by month
• Report longest missing streaks for each station and variable
• Provide descriptive statistics for selected variables
• Run in Jupyter or from the command line
• Keep all outputs in memory by default
• Save output files only when the user requests it

Planned direction

This package is designed to grow beyond EDA. Future releases may include tools for climate data download, cleaning, spatial aggregation, trend analysis, drought analysis, precipitation indices, temperature extremes, and visualization.

Installation

After the package is published to PyPI:

pip install climate-indepth-analysis

For local development from this folder:

pip install -e .

Jupyter usage

from climate_indepth_analysis import run_eda

results = run_eda(
    input_path="my_climate_data.csv",
    station_col="STATION",
    date_col="DATE",
    needed_cols=["PRCP", "TMAX", "TMIN"],
    start_date="1980-01-01",
    end_date="2025-12-31",
    save_outputs=False,
)

results["overall"]
results["station_summary"].head()
results["monthly"].head()

Save outputs only when needed

By default, the package does not save any file. To save summary tables and a text report, set save_outputs=True.

results = run_eda(
    input_path="my_climate_data.csv",
    station_col="ID",
    date_col="date",
    save_outputs=True,
    output_dir="my_climate_eda_results",
)

The full station-date calendar can be very large. It is not saved unless you also set save_full_calendar=True.

results = run_eda(
    input_path="my_climate_data.csv",
    save_outputs=True,
    output_dir="my_climate_eda_results",
    save_full_calendar=True,
)

Command-line usage

Run without saving files:

climate-indepth-analysis --input my_climate_data.csv

Run with custom station and date columns:

climate-indepth-analysis --input my_climate_data.csv --station-col ID --date-col date

Save output files to a custom folder:

climate-indepth-analysis --input my_climate_data.csv --save-outputs --output-dir my_climate_eda_results

You can also use the shorter command:

climate-eda --input my_climate_data.csv

Default settings

station_col = "STATION"
date_col = "DATE"
needed_cols = ["PRCP", "TMAX", "TMIN"]
start_date = "1980-01-01"
end_date = "2025-12-31"
save_outputs = False
output_dir = "climate_indepth_analysis_output"

Output dictionary

run_eda() returns a dictionary with these tables:

results["diagnostics"]
results["inventory"]
results["overall"]
results["station_summary"]
results["monthly"]
results["variable_stats"]
results["full_calendar"]

Files saved when save_outputs=True

00_cleanliness_report.txt
01_file_calendar_diagnostics.csv
02_station_inventory_coverage.csv
03_overall_missing_summary.csv
04_station_missing_summary.csv
05_monthly_missing_summary_long.csv
06_variable_descriptive_stats.csv

If save_full_calendar=True, the package also saves:

07_full_station_date_calendar.parquet

If Parquet support is unavailable, it falls back to CSV.

Build and upload to PyPI

Install build tools:

python -m pip install --upgrade build twine

Build the package:

python -m build

Check the package:

python -m twine check dist/*

Upload to TestPyPI first:

python -m twine upload --repository testpypi dist/*

Upload to PyPI:

python -m twine upload dist/*

License

MIT License.