Geo Daily Tools

A small Python project for reusable geospatial/data-science daily utilities: initial dataset inspection, pandas validation, geospatial QA/QC, model-ready feature preparation, and interview-style practice scripts.

Quick start

python -m venv .venv
# Windows PowerShell
.venv\Scripts\Activate.ps1
# macOS/Linux
source .venv/bin/activate

pip install -e .[dev,io]

Run the examples:

python examples/01_clean_sensor_records.py
python examples/02_geodataframe_workflow.py
python examples/03_modeling_prep_workflow.py

Run tests:

pytest

Project layout

geo-daily-tools/
├─ pyproject.toml
├─ README.md
├─ examples/
│  ├─ 01_clean_sensor_records.py
│  ├─ 02_geodataframe_workflow.py
│  └─ 03_modeling_prep_workflow.py
├─ src/
│  └─ geo_daily_tools/
│     ├─ __init__.py
│     ├─ inspection.py
│     ├─ validation.py
│     ├─ geo_validation.py
│     ├─ modeling.py
│     ├─ io_utils.py
│     └─ sample_data.py
└─ tests/
   ├─ test_validation.py
   ├─ test_geo_validation.py
   └─ test_modeling.py

Interview/coding patterns this project is built around

1. Initial inspection

Use quick_inspect(df) to print shape, columns, dtypes, missingness, numeric summaries, and categorical summaries.

2. Drop-reason validation

Use validate_sensor_records(records) to create a cleaned dataframe and a drop summary. The function uses a drop_reason column so invalid rows are auditable.

3. GeoDataFrame creation

Use points_from_latlon(df) to create an EPSG:4326 GeoDataFrame from lat/lon columns.

4. Spatial modeling prep

Use prepare_model_inputs(df, features, target) and group_train_test_split(...) to avoid accidental spatial leakage.

Notes

This is intentionally simple and readable. It is built for reusable daily work and live-coding practice, not as a heavyweight framework.