Skip to content
Switch branches/tags
Go to file


hydropandas Codacy Badge Codacy Badge Documentation Status PyPi Binder


The hydropandas module is a Python package for reading timeseries data into DataFrames.

The Hydropandas package allows users to manipulate data using all of the wonderful features included in pandas extented with custom methods and attributes related to the timeseries. The hydropandas module extends pandas.DataFrame with extra functionality and stores metadata related to the type of measurements.


Install the module with pip:

pip install hydropandas

Hydropandas requires numpy, scipy, matplotlib, pandas, geopandas, requests and zeep.

For some functionality additional packages are required:

  • pastastore: for reading or storing data from PastaStore
  • bokeh, branca, folium: for interactive maps
  • flopy: for reading data from MODFLOW models
  • xarray: for loading data from REGIS

For installing in development mode, clone the repository and install by typing pip install -e . from the module root directory.

If you have trouble installing hydropandas, refer to the Dependencies section below.

Example usage

Importing a single DINO csv file:

import hydropandas as hpd
fname = './tests/data/2019-Dino-test/Grondwaterstanden_Put/B33F0080001_1.csv'
gw = hpd.GroundwaterObs.from_dino(fname=fname, verbose=True)

Or for a zipfile:

import hydropandas as hpd
dinozip = './tests/data/2019-Dino-test/'
dino_gw = hpd.ObsCollection.from_dino(dirname=dinozip,

The Obs class

The Obs class holds the measurements and metadata for one timeseries. There are currently 5 specific Obs classes for different types of measurements:

  • GroundwaterObs: for groundwater measurements
  • GroundwaterQualityObs: for groundwater quality measurements
  • WaterlvlObs: for surface water level measurements
  • ModelObs: for hydropandas from a MODFLOW model
  • KnmiObs: for (daily) KNMI hydropandas

Each of these Obs classes is essentially a pandas DataFrame with additional methods and attributes related to the type of measurement that it holds. The classes also contain specific methods to read data from specific sources.

The ObsCollection class

The ObsCollection class, as the name implies, represents a collection of Obs classes, e.g. 10 timeseries of the groundwater level in a certain area. The ObsCollection is also a pandas DataFrame in which each timeseries is stored in a different row. Each row contains metadata (e.g. latitude and longitude of the observation point) and the Obs object (DataFrame) that holds the measurements. It is recommended to let an ObsCollection contain only one Obs type, e.g. to create an ObsCollection for 10 GroundwaterObs, and a separate ObsCollection for 5 KnmiObs.

Like the Obs class, the ObsCollection class contains a bunch of methods for reading data from different sources. See the next section for supported data sources.

Supported data sources

Currently supported datasources that can be read:

  • DINO csv
  • WISKI csv
  • Artesia Fieldlogger for Android and iOS
  • Pastas projects (deprecated)
  • Pastastore, for managing Pastas timeseries and models
  • PyStore, a fast datastore for pandas timeseries
  • Arctic, a timeseries / dataframe database that sits atop MongoDB
  • KNMI data
  • MODFLOW groundwater models
  • IMOD groundwater models

ObsCollection can be exported to:

  • Artesia Fieldlogger
  • Shapefile
  • Pastas projects (deprecated)
  • Pastastore
  • Arctic
  • Pystore


Hydropandas (indirectly) uses some packages that cannot be installed automatically with pip on Windows. These packages are:

  • GDAL
  • Fiona
  • Shapely
  • Python-snappy
  • Fastparquet

If you do not have these packages already it is recommended to first try installing them with conda install <pkg>. Otherwise, read the instructions below how to install them manually.

Download the packages from Christoph Gohlke's website. Use CTRL+F to find the download link on the page. Be sure to download the correct version of the package. The Python version should match your Python version. Also the architecture should match (i.e. 64bits vs 32bits). For example:

  • GDAL-3.1.4-cp38-cp38-win_amd64.whl

This is the GDAL version for Python 3.8 (as can be seen from the cp38 in the name), for 64-bits Python (as derived from the amd64 in the name).

Once you have downloaded the correct files, navigate to the directory in which you saved your downloads. Now type the following commands (the order is important):

  1. pip install GDAL-3.1.4-cp38-cp38-win_amd64.whl
  2. pip install Fiona-1.8.17-cp38-cp38-win_amd64.whl
  3. pip install Shapely-1.7.1-cp38-cp38-win_amd64.whl
  4. pip install python_snappy-0.5.4-cp38-cp38-win_amd64.whl
  5. pip install fastparquet-0.4.1-cp38-cp38-win_amd64.whl

After you've done this you can install hydropandas using pip install hydropandas.


  • Onno Ebbens, Artesia
  • Ruben Caljé, Artesia
  • Davíd Brakenhoff, Artesia