Datasets:
metadata
license: apache-2.0
language:
- en
tags:
- groundwater
- india
- groundwater level
- wells data
- groundwater india
pretty_name: Groundwater Level Data of India
size_categories:
- 10K<n<100K
India Groundwater Level Dataset
Dataset Description
This dataset contains comprehensive groundwater level measurements from monitoring stations across India, compiled from the India Water Resources Information System (India-WRIS). The data has been processed and structured in a time-series format suitable for machine learning applications, hydrological analysis, and water resource management studies.
The dataset aggregates measurements from groundwater monitoring wells operated by the Central Ground Water Board (CGWB) across multiple states and union territories of India, spanning several decades of observations. Each row represents a unique monitoring station with its metadata and monthly groundwater level readings.
Key Features
- Temporal Coverage: Historical groundwater measurements spanning from 1970s to 2025
- Spatial Coverage: Pan-India coverage across multiple states and districts
- Data Quality: Filtered to include only stations with ≥20 data points for reliable analysis
- Temporal Resolution: Monthly groundwater level measurements
- Standardized Format: Cleaned and pivoted structure with chronologically ordered time-series data
Source
Data sourced from India Water Resources Information System (India-WRIS), maintained by the Central Ground Water Board (CGWB), Ministry of Jal Shakti, Government of India.
Uses
Primary Applications
Groundwater Trend Analysis
- Long-term groundwater depletion studies
- Seasonal variation analysis
- Aquifer behavior characterization
- Climate change impact assessment
Time Series Forecasting
- Predicting future groundwater levels
- Water availability forecasting
- Early warning systems for groundwater stress
- Drought prediction and monitoring
Spatial Analysis
- Regional groundwater mapping
- Identifying critical zones
- Inter-district/inter-state comparisons
- Aquifer connectivity studies
Machine Learning Applications
- Training models for groundwater level prediction
- Anomaly detection in groundwater patterns
- Classification of aquifer types
- Feature engineering for hydrological models
Policy and Planning
- Water resource management decisions
- Agricultural planning and irrigation scheduling
- Urban water supply planning
- Sustainable groundwater extraction policies
Research Areas
- Hydrogeology and hydroinformatics
- Environmental science and climate studies
- Agricultural water management
- Urban planning and sustainability
- Data science and predictive modeling
Direct Use
Loading the Dataset
import pandas as pd
from datasets import load_dataset
# Load from HuggingFace
dataset = load_dataset("Daksh17440/india_gwl")
df = dataset['train'].to_pandas()
# Or load directly from CSV
df = pd.read_csv("combined_pivoted_min40pts.csv")
Basic Analysis
# View dataset structure
print(f"Dataset shape: {df.shape}")
print(f"Number of stations: {len(df)}")
# Access metadata columns
metadata_cols = ['stationCode', 'stationName', 'stationType', 'latitude',
'longitude', 'district', 'state', 'tehsil', 'block',
'village', 'unit', 'well_type', 'well_depth', 'well_aquifer_type']
# Access time-series data (columns with year-month format)
date_cols = [col for col in df.columns if any(char.isdigit() for char in str(col))]
# Filter stations by location
delhi_stations = df[df['state'] == 'Delhi']
# Get groundwater levels for a specific station
station_data = df[df['stationCode'] == '2.33E+14'][date_cols]
Time Series Analysis Example
import matplotlib.pyplot as plt
# Extract time series for a station
station = df.iloc[0]
time_series = station[date_cols].dropna()
# Convert to datetime index
dates = pd.to_datetime([col for col in time_series.index])
values = time_series.values
# Plot
plt.figure(figsize=(12, 6))
plt.plot(dates, values)
plt.title(f"Groundwater Level - {station['stationName']}")
plt.xlabel("Date")
plt.ylabel("Groundwater Level (m)")
plt.grid(True)
plt.show()
Spatial Analysis Example
import geopandas as gpd
import matplotlib.pyplot as plt
# Create GeoDataFrame
gdf = gpd.GeoDataFrame(
df,
geometry=gpd.points_from_xy(df.longitude, df.latitude),
crs="EPSG:4326"
)
# Calculate mean groundwater level per station
df['mean_gwl'] = df[date_cols].mean(axis=1)
# Plot spatial distribution
fig, ax = plt.subplots(figsize=(12, 10))
gdf.plot(column='mean_gwl', cmap='RdYlBu', legend=True, ax=ax,
markersize=20, alpha=0.6)
plt.title("Mean Groundwater Levels Across India")
plt.show()
Machine Learning Example
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestRegressor
# Prepare data for forecasting (predict next month)
# Using previous 12 months to predict the 13th month
def prepare_ml_data(df, lookback=12):
X, y = [], []
for idx, row in df.iterrows():
ts = row[date_cols].dropna()
if len(ts) >= lookback + 1:
for i in range(len(ts) - lookback):
X.append(ts.iloc[i:i+lookback].values)
y.append(ts.iloc[i+lookback])
return np.array(X), np.array(y)
X, y = prepare_ml_data(df)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
# Train model
model = RandomForestRegressor(n_estimators=100)
model.fit(X_train, y_train)
# Evaluate
score = model.score(X_test, y_test)
print(f"Model R² Score: {score:.3f}")
Dataset Structure
Data Format
The dataset is provided in CSV format with the following structure:
- Rows: Each row represents a unique groundwater monitoring station
- Columns: Metadata columns followed by time-series columns
Column Types
1. Metadata Columns (Fixed)
| Column Name | Type | Description |
|---|---|---|
stationCode |
String/Numeric | Unique identifier for the monitoring station |
stationName |
String | Name/location description of the station |
stationType |
String | Type of monitoring station (e.g., "Ground Water") |
latitude |
Float | Latitude coordinate (decimal degrees) |
longitude |
Float | Longitude coordinate (decimal degrees) |
district |
String | District name where station is located |
state |
String | State/UT name where station is located |
tehsil |
String | Tehsil/Taluka name (administrative subdivision) |
block |
String | Block name (administrative subdivision) |
village |
String | Village name where station is located |
unit |
String | Unit of measurement for groundwater level |
well_type |
String | Type of well (e.g., "Dug Well", "Bore Well") |
well_depth |
Float | Depth of the well (in meters) |
well_aquifer_type |
String | Type of aquifer (e.g., "Unconfined", "Confined") |
2. Time-Series Columns (Variable)
- Format:
YYYY-MM(e.g., "2020-01", "2021-08") - Type: Float
- Description: Groundwater level measurement for that month
- Unit: Meters below ground level (mbgl) or as specified in 'unit' column
- Missing Values: Empty cells indicate no measurement for that period
- Chronological Order: Columns are sorted chronologically from earliest to latest date
Data Characteristics
- Number of Rows: Varies based on filtering criteria (stations with ≥40 data points)
- Number of Columns: ~500-1000+ (14 metadata + varying time-series columns)
- Missing Data: Present in time-series columns due to irregular measurement schedules
- Date Range: Typically spans 1970-2025, varies by station
- Spatial Coverage: All major states and union territories of India
- Coordinate System: WGS84 (EPSG:4326)
Data Processing Pipeline
- Source Data Extraction: Downloaded from India-WRIS API
- Format Standardization: Irregular CSV rows corrected (24-25 column rows normalized to 23)
- Date Parsing: Measurement dates extracted and validated
- Pivoting: Transformed from long format to wide format (time-series structure)
- Filtering: Retained only stations with ≥40 valid measurements
- Chronological Sorting: Time columns ordered from earliest to latest
- Quality Checks: Validated coordinates, removed duplicates
Data Quality Notes
- Coordinate Validation: Latitude (8°-37°N) and Longitude (68°-98°E) validated for India
- Missing Values: Handled with NA, not imputed to preserve data integrity
- Temporal Gaps: Common due to operational/seasonal monitoring schedules
- Spatial Distribution: Coverage varies by state; some regions more densely monitored
- Measurement Frequency: Primarily monthly, but can vary by station
Example Row Structure
stationCode: 2.33E+14
stationName: Amarpur Dhawajina
stationType: Ground Water
latitude: 23.525
longitude: 91.6917
district: SOUTH TRIPURA
state: Tripura
...
2020-01: 9.5
2020-11: (null)
2021-01: 7.58
2021-08: 2.75
...
Citation
If you use this dataset in your research, please cite:
@dataset{india_groundwater_2025,
title={India Groundwater Level Dataset},
author={Daksh Gogna},
year={2025},
publisher={HuggingFace},
url={https://huggingface.co/datasets/Daksh17440/india_gwl},
note={Data sourced from India Water Resources Information System (India-WRIS),
Central Ground Water Board (CGWB), Ministry of Jal Shakti, Government of India}
}
License
This dataset is derived from publicly available data from India-WRIS (Government of India). Please refer to the original source for specific usage terms and conditions.
Limitations and Considerations
- Temporal Coverage: Not all stations have continuous measurements across the entire time span
- Spatial Bias: Monitoring density varies across regions; urban and agricultural areas may have better coverage
- Data Gaps: Missing values are common due to operational constraints, seasonal access issues, or equipment failures
- Measurement Standardization: Different well types and measurement methods may introduce variability
- Aquifer Complexity: Single point measurements may not fully represent regional aquifer conditions
- Climate Influence: Seasonal monsoons significantly impact groundwater levels in India
- Human Impact: Groundwater extraction for agriculture and urban use affects natural trends
Contact and Contributions
For questions, issues, or contributions, please open an issue or contact daksh_g@es.iitr.ac.in.
Acknowledgments
- Central Ground Water Board (CGWB), Ministry of Jal Shakti, Government of India
- India Water Resources Information System (India-WRIS)
- All field teams involved in groundwater monitoring across India