diff --git a/.github/workflows/CI.yml b/.github/workflows/CI.yml
index 91f42577d..d01bba6be 100644
--- a/.github/workflows/CI.yml
+++ b/.github/workflows/CI.yml
@@ -70,34 +70,32 @@ jobs:
       fail-fast: false
 
     steps:
-      # used to reset cache every month
-      - name: Get current year-month
-        id: date
-        run: echo "date=$(date +'%Y-%m')" >> $GITHUB_OUTPUT
-      
-      - name: Get pip cache dir
-        id: pip-cache
-        run: |
-          echo "dir=$(pip cache dir)" >> $GITHUB_OUTPUT
-
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4.1.6
 
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5.1.0
         with:
           python-version: ${{ matrix.python-version }}
+          check-latest: true
           cache: 'pip'
-          cache-dependency-path: '**/requirements.txt'
+          cache-dependency-path: |
+            **/requirements.txt
+            **/requirements-extras.txt
+            **/requirements-tests.txt
 
-      - name: Cache test_env
-        uses: actions/cache@v3
+      - name: Get current hash (SHA) of the elephant_data repo
+        id: elephant-data
+        run: |
+            echo "dataset_hash=$(git ls-remote https://gin.g-node.org/NeuralEnsemble/elephant-data.git HEAD | cut -f1)" >> $GITHUB_OUTPUT
+
+      - uses: actions/cache/restore@v4.0.2
+        # Loading cache of elephant-data
+        id: cache-datasets
         with:
-          path: ${{ steps.pip-cache.outputs.dir }}
-          # Look to see if there is a cache hit for the corresponding requirements files
-          # cache will be reset on changes to any requirements or every month
-          key: ${{ runner.os }}-venv-${{ hashFiles('**/requirements.txt') }}-${{ hashFiles('**/requirements-tests.txt') }}
-            -${{ hashFiles('**/requirements-extras.txt') }}-${{ hashFiles('**/CI.yml') }}-${{ hashFiles('setup.py') }}
-            -${{ steps.date.outputs.date }}
+          path: ~/elephant-data
+          key: datasets-${{ steps.elephant-data.outputs.dataset_hash }}
+          restore-keys: datasets-
+          enableCrossOsArchive: true 
 
       - name: Install dependencies
         run: |
@@ -112,6 +110,11 @@ jobs:
 
       - name: Test with pytest
         run: |
+          if [ -d ~/elephant-data ]; then
+          export ELEPHANT_DATA_LOCATION=~/elephant-data
+          echo $ELEPHANT_DATA_LOCATION
+          fi
+      
           coverage run --source=elephant -m pytest
           coveralls --service=github || echo "Coveralls submission failed"
         env:
@@ -146,6 +149,19 @@ jobs:
           path: ~/conda_pkgs_dir
           key: ${{ runner.os }}-conda-${{hashFiles('requirements/environment.yml') }}-${{ hashFiles('**/CI.yml') }}-${{ steps.date.outputs.date }}
 
+      - name: Get current hash (SHA) of the elephant_data repo
+        id: elephant-data
+        run: |
+            echo "dataset_hash=$(git ls-remote https://gin.g-node.org/NeuralEnsemble/elephant-data.git HEAD | cut -f1)" >> $GITHUB_OUTPUT
+
+      - uses: actions/cache/restore@v4.0.2
+        # Loading cache of elephant-data
+        id: cache-datasets
+        with:
+          path: ~/elephant-data
+          key: datasets-${{ steps.elephant-data.outputs.dataset_hash }}
+          restore-keys: datasets-      
+
       - uses: conda-incubator/setup-miniconda@a4260408e20b96e80095f42ff7f1a15b27dd94ca # corresponds to v3.0.4
         with:
           auto-update-conda: true
@@ -173,6 +189,10 @@ jobs:
       - name: Test with pytest
         shell: bash -l {0}
         run: |
+          if [ -d ~/elephant-data ]; then
+          export ELEPHANT_DATA_LOCATION=~/elephant-data
+          echo $ELEPHANT_DATA_LOCATION
+          fi
           pytest --cov=elephant
 
   # __        ___           _
@@ -192,24 +212,32 @@ jobs:
         os: [windows-latest]
 
     steps:
-      - name: Get current year-month
-        id: date
-        run: echo "date=$(date +'%Y-%m')" >> $GITHUB_OUTPUT
-
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4.1.6
 
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5.1.0
         with:
           python-version: ${{ matrix.python-version }}
+          check-latest: true
+          cache: 'pip'
+          cache-dependency-path: |
+            **/requirements.txt
+            **/requirements-extras.txt
+            **/requirements-tests.txt
 
-      - name: Cache pip
-        uses: actions/cache@v3
+      - name: Get current hash (SHA) of the elephant_data repo
+        id: elephant-data
+        run: |
+            echo "dataset_hash=$(git ls-remote https://gin.g-node.org/NeuralEnsemble/elephant-data.git HEAD | cut -f1)" >> $GITHUB_OUTPUT
+
+      - uses: actions/cache/restore@v4.0.2
+        # Loading cache of elephant-data
+        id: cache-datasets
         with:
-          path: ~\AppData\Local\pip\Cache
-          # Look to see if there is a cache hit for the corresponding requirements files
-          key: ${{ runner.os }}-pip-${{ hashFiles('**/requirements.txt') }}-${{ hashFiles('**/requirements-tests.txt') }}
-            -${{ hashFiles('**/requirements-extras.txt') }}-${{ hashFiles('setup.py') }} -${{ hashFiles('**/CI.yml') }}-${{ steps.date.outputs.date }}
+          path: ~/elephant-data
+          key: datasets-${{ steps.elephant-data.outputs.dataset_hash }}
+          restore-keys: datasets-
+          enableCrossOsArchive: true    
 
       - name: Install dependencies
         run: |
@@ -224,6 +252,10 @@ jobs:
 
       - name: Test with pytest
         run: |
+          if (Test-Path "$env:USERPROFILE\elephant-data") {
+              $env:ELEPHANT_DATA_LOCATION = "$env:USERPROFILE\elephant-data"
+              Write-Output $env:ELEPHANT_DATA_LOCATION
+          }
           pytest --cov=elephant
 
   #  __  __ ____ ___
@@ -246,29 +278,32 @@ jobs:
       fail-fast: false
 
     steps:
-      - name: Get current year-month
-        id: date
-        run: echo "date=$(date +'%Y-%m')" >> $GITHUB_OUTPUT
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4.1.6
 
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5.1.0
         with:
           python-version: ${{ matrix.python-version }}
+          check-latest: true
+          cache: 'pip'
+          cache-dependency-path: |
+            **/requirements.txt
+            **/requirements-extras.txt
+            **/requirements-tests.txt
 
-      - name: Get pip cache dir
-        id: pip-cache
+      - name: Get current hash (SHA) of the elephant_data repo
+        id: elephant-data
         run: |
-          echo "dir=$(pip cache dir)" >> $GITHUB_OUTPUT
+            echo "dataset_hash=$(git ls-remote https://gin.g-node.org/NeuralEnsemble/elephant-data.git HEAD | cut -f1)" >> $GITHUB_OUTPUT
 
-      - name: Cache test_env
-        uses: actions/cache@v3
+      - uses: actions/cache/restore@v4.0.2
+        # Loading cache of elephant-data
+        id: cache-datasets
         with:
-          path: ${{ steps.pip-cache.outputs.dir }}
-          # look to see if there is a cache hit for the corresponding requirements files
-          # cache will be reset on changes to any requirements or every month
-          key: ${{ runner.os }}-venv-${{ hashFiles('**/requirements.txt') }}-${{ hashFiles('**/requirements-tests.txt') }}
-            -${{ hashFiles('**/requirements-extras.txt') }}-${{ hashFiles('setup.py') }} -${{ hashFiles('**/CI.yml') }}-${{ steps.date.outputs.date }}
+          path: ~/elephant-data
+          key: datasets-${{ steps.elephant-data.outputs.dataset_hash }}
+          restore-keys: datasets-
+          enableCrossOsArchive: true 
 
       - name: Setup environment
         run: |
@@ -287,6 +322,10 @@ jobs:
 
       - name: Test with pytest
         run: |
+          if [ -d ~/elephant-data ]; then
+          export ELEPHANT_DATA_LOCATION=~/elephant-data
+          echo $ELEPHANT_DATA_LOCATION
+          fi
           mpiexec -n 1 python -m mpi4py -m coverage run --source=elephant -m pytest
           coveralls --service=github || echo "Coveralls submission failed"
         env:
@@ -316,7 +355,7 @@ jobs:
         id: date
         run: echo "date=$(date +'%Y-%m')" >> $GITHUB_OUTPUT
 
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4.1.6
 
       - name: Get pip cache dir
         id: pip-cache
@@ -330,6 +369,20 @@ jobs:
 
           key: ${{ runner.os }}-pip-${{hashFiles('requirements/environment-tests.yml') }}-${{ hashFiles('**/CI.yml') }}-${{ steps.date.outputs.date }}
 
+      - name: Get current hash (SHA) of the elephant_data repo
+        id: elephant-data
+        run: |
+            echo "dataset_hash=$(git ls-remote https://gin.g-node.org/NeuralEnsemble/elephant-data.git HEAD | cut -f1)" >> $GITHUB_OUTPUT
+
+      - uses: actions/cache/restore@v4.0.2
+        # Loading cache of elephant-data
+        id: cache-datasets
+        with:
+          path: ~/elephant-data
+          key: datasets-${{ steps.elephant-data.outputs.dataset_hash }}
+          restore-keys: datasets-
+          enableCrossOsArchive: true 
+
       - uses: conda-incubator/setup-miniconda@030178870c779d9e5e1b4e563269f3aa69b04081 # corresponds to v3.0.3
         with:
           auto-update-conda: true
@@ -358,6 +411,10 @@ jobs:
       - name: Test with pytest
         shell: bash -el {0}
         run: |
+          if [ -d ~/elephant-data ]; then
+          export ELEPHANT_DATA_LOCATION=~/elephant-data
+          echo $ELEPHANT_DATA_LOCATION
+          fi
           pytest --cov=elephant
 
   #  ____
@@ -383,7 +440,7 @@ jobs:
         id: date
         run: echo "date=$(date +'%Y-%m')" >> $GITHUB_OUTPUT
 
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4.1.6
 
       - name: Get pip cache dir
         id: pip-cache
@@ -448,10 +505,10 @@ jobs:
       - name: Get current year-month
         id: date
         run: echo "::set-output name=date::$(date +'%Y-%m')"
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4.1.6
 
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5.1.0
         with:
           python-version: ${{ matrix.python-version }}
 
diff --git a/.github/workflows/cache_elephant_data.yml b/.github/workflows/cache_elephant_data.yml
new file mode 100644
index 000000000..ab17af4f9
--- /dev/null
+++ b/.github/workflows/cache_elephant_data.yml
@@ -0,0 +1,65 @@
+name: Create caches for elephant_data
+
+on:
+  workflow_dispatch: # Workflow can be triggered manually via GH actions webinterface
+  push:  # When something is pushed into master this checks if caches need to re-created
+    branches:
+      - master
+  schedule:
+    - cron: "11 23 * * *"  # Daily at 23:11 UTC
+
+
+jobs:
+  create-data-cache-if-missing:
+    name: Caching data env
+    runs-on: ubuntu-latest
+    strategy:
+      # do not cancel all in-progress jobs if any matrix job fails
+      fail-fast: false
+
+    steps:
+      - name: Get current hash (SHA) of the elephant_data repo
+        id: elephant-data
+        run: |
+            echo "dataset_hash=$(git ls-remote https://gin.g-node.org/NeuralEnsemble/elephant-data.git HEAD | cut -f1)" >> $GITHUB_OUTPUT
+
+      - uses: actions/cache@v4.0.2
+        # Loading cache of elephant-data
+        id: cache-datasets
+        with:
+          path: ~/elephant-data
+          key: datasets-${{ steps.elephant-data.outputs.dataset_hash }}
+
+      - name: Cache found?
+        run: echo "Cache-hit == ${{steps.cache-datasets.outputs.cache-hit == 'true'}}"
+
+      - name: Configuring git
+        if: steps.cache-datasets.outputs.cache-hit != 'true'
+        run: |
+            git config --global user.email "elephant_ci@fake_mail.com"
+            git config --global user.name "elephant CI"
+            git config --global filter.annex.process "git-annex filter-process"  # recommended for efficiency
+
+      - name: Install Datalad Linux
+        if: steps.cache-datasets.outputs.cache-hit != 'true'
+        run: |
+          python -m pip install -U pip  # Official recommended way
+          pip install datalad-installer
+          datalad-installer --sudo ok git-annex --method datalad/packages
+          pip install datalad
+                
+      - name: Download dataset
+        id: download-dataset
+        if: steps.cache-datasets.outputs.cache-hit != 'true'
+        # Download repository and also fetch data
+        run: |
+          cd ~
+          datalad --version
+          datalad install --recursive --get-data https://gin.g-node.org/NeuralEnsemble/elephant-data
+
+      - name: Show size of the cache to assert data is downloaded
+        run: |
+          cd ~
+          du -hs ~/elephant-data
+          ls -lh ~/elephant-data
+
diff --git a/elephant/datasets.py b/elephant/datasets.py
index 58d52d31f..39ed56911 100644
--- a/elephant/datasets.py
+++ b/elephant/datasets.py
@@ -1,17 +1,19 @@
 import hashlib
+import os
+import ssl
 import tempfile
+from urllib.parse import urlparse
 import warnings
-import ssl
-
-from elephant import _get_version
+from os import environ, getenv
 from pathlib import Path
-from urllib.request import urlretrieve, urlopen
 from urllib.error import HTTPError, URLError
+from urllib.request import urlopen, urlretrieve
 from zipfile import ZipFile
-from os import environ, getenv
 
 from tqdm import tqdm
 
+from elephant import _get_version
+
 ELEPHANT_TMP_DIR = Path(tempfile.gettempdir()) / "elephant"
 
 
@@ -74,65 +76,76 @@ def download(url, filepath=None, checksum=None, verbose=True):
 def download_datasets(repo_path, filepath=None, checksum=None,
                       verbose=True):
     r"""
-        This function can be used to download files from elephant-data using
-        only the path relative to the root of the elephant-data repository.
-        The default URL used, points to elephants corresponding release of
-        elephant-data.
-        Different versions of the elephant package may require different
-        versions of elephant-data.
-        e.g. the follwoing URLs:
-        -  https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/0.0.1
-           points to release v0.0.1.
-        -  https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/master
-           always points to the latest state of elephant-data.
-        -  https://datasets.python-elephant.org/
-           points to the root of elephant data
-
-        To change this URL, use the environment variable `ELEPHANT_DATA_URL`.
-        When using data, which is not yet contained in the master branch or a
-        release of elephant data, e.g. during development, this variable can
-        be used to change the default URL.
-        For example to use data on branch `multitaper`, change the
-        `ELEPHANT_DATA_URL` to
-        https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/multitaper.
-        For a complete example, see Examples section.
-
-        Parameters
-        ----------
-        repo_path : str
-            String denoting the path relative to elephant-data repository root
-        filepath : str, optional
-            Path to temporary folder where the downloaded files will be stored
-        checksum : str, otpional
-            Checksum to verify dara integrity after download
-        verbose : bool, optional
-            Whether to disable the entire progressbar wrapper [].
-            If set to None, disable on non-TTY.
-            Default: True
-
-        Returns
-        -------
-        filepath : pathlib.Path
-            Path to downloaded files.
-
-
-        Notes
-        -----
-        The default URL always points to elephant-data. Please
-        do not change its value. For development purposes use the environment
-        variable 'ELEPHANT_DATA_URL'.
-
-        Examples
-        --------
-        The following example downloads a file from elephant-data branch
-        'multitaper', by setting the environment variable to the branch URL:
-
-        >>> import os
-        >>> from elephant.datasets import download_datasets
-        >>> os.environ["ELEPHANT_DATA_URL"] = "https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/multitaper" # noqa
-        >>> download_datasets("unittest/spectral/multitaper_psd/data/time_series.npy") # doctest: +SKIP
-        PosixPath('/tmp/elephant/time_series.npy')
-        """
+    This function can be used to download files from elephant-data using
+    only the path relative to the root of the elephant-data repository.
+    The default URL used, points to elephants corresponding release of
+    elephant-data.
+    Different versions of the elephant package may require different
+    versions of elephant-data.
+    e.g. the following URLs:
+    -  https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/0.0.1
+       points to release v0.0.1.
+    -  https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/master
+       always points to the latest state of elephant-data.
+    -  https://datasets.python-elephant.org/
+       points to the root of elephant data
+
+    To change this URL, use the environment variable `ELEPHANT_DATA_LOCATION`.
+    When using data, which is not yet contained in the master branch or a
+    release of elephant data, e.g. during development, this variable can
+    be used to change the default URL.
+    For example to use data on branch `multitaper`, change the
+    `ELEPHANT_DATA_LOCATION` to
+    https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/multitaper.
+    For a complete example, see Examples section.
+        
+    To use a local copy of elephant-data, use the environment variable
+    `ELEPHANT_DATA_LOCATION`, e.g. set to /home/user/elephant-data.
+        
+    Parameters
+    ----------
+    repo_path : str
+        String denoting the path relative to elephant-data repository root
+    filepath : str, optional
+        Path to temporary folder where the downloaded files will be stored
+    checksum : str, optional
+        Checksum to verify data integrity after download
+    verbose : bool, optional
+        Whether to disable the entire progressbar wrapper [].
+        If set to None, disable on non-TTY.
+        Default: True
+
+    Returns
+    -------
+    filepath : pathlib.Path
+        Path to downloaded files.
+
+
+    Notes
+    -----
+    The default URL always points to elephant-data. Please
+    do not change its value. For development purposes use the environment
+    variable 'ELEPHANT_DATA_LOCATION'.
+
+    Examples
+    --------
+    The following example downloads a file from elephant-data branch
+    'multitaper', by setting the environment variable to the branch URL:
+
+    >>> import os
+    >>> from elephant.datasets import download_datasets
+    >>> os.environ["ELEPHANT_DATA_LOCATION"] = "https://web.gin.g-node.org/NeuralEnsemble/elephant-data/raw/multitaper" # noqa
+    >>> download_datasets("unittest/spectral/multitaper_psd/data/time_series.npy") # doctest: +SKIP
+    PosixPath('/tmp/elephant/time_series.npy')
+    """
+
+    env_var = 'ELEPHANT_DATA_LOCATION'
+    if env_var in os.environ:  # user did set path or URL
+        if os.path.exists(getenv(env_var)):
+            return Path(f"{getenv(env_var)}/{repo_path}")
+        elif urlparse(getenv(env_var)).scheme not in ('http', 'https'):
+            raise ValueError(f"The environment variable {env_var} must be set to either an existing file system path "
+                             f"or a valid URL. Given value: '{getenv(env_var)}' is neither.")
 
     # this url redirects to the current location of elephant-data
     url_to_root = "https://datasets.python-elephant.org/"
@@ -141,7 +154,7 @@ def download_datasets(repo_path, filepath=None, checksum=None,
     # (version elephant is equal to version elephant-data)
     default_url = url_to_root + f"raw/v{_get_version()}"
 
-    if 'ELEPHANT_DATA_URL' not in environ:  # user did not set URL
+    if env_var not in environ:  # user did not set URL
         # is 'version-URL' available? (not for elephant development version)
         try:
             urlopen(default_url+'/README.md')
@@ -149,7 +162,7 @@ def download_datasets(repo_path, filepath=None, checksum=None,
         except HTTPError as error:
             # if corresponding elephant-data version is not found,
             # use latest commit of elephant-data
-            default_url = url_to_root + f"raw/master"
+            default_url = url_to_root + "raw/master"
 
             warnings.warn(f"No corresponding version of elephant-data found.\n"
                           f"Elephant version: {_get_version()}. "
@@ -164,12 +177,12 @@ def download_datasets(repo_path, filepath=None, checksum=None,
                 ctx.check_hostname = True
                 urlopen(default_url + '/README.md')
             except HTTPError:  # e.g. 404
-                default_url = url_to_root + f"raw/master"
+                default_url = url_to_root + "raw/master"
 
             warnings.warn(f"Data URL:{default_url}, error: {error}."
                           f"{error.reason}")
 
-    url = f"{getenv('ELEPHANT_DATA_URL', default_url)}/{repo_path}"
+    url = f"{getenv(env_var, default_url)}/{repo_path}"
 
     return download(url, filepath, checksum, verbose)
 
diff --git a/elephant/test/test_causality.py b/elephant/test/test_causality.py
index 552fe2f05..648af2a04 100644
--- a/elephant/test/test_causality.py
+++ b/elephant/test/test_causality.py
@@ -16,7 +16,7 @@
 
 from elephant.spectral import multitaper_cross_spectrum, multitaper_coherence
 import elephant.causality.granger
-from elephant.datasets import download_datasets, ELEPHANT_TMP_DIR
+from elephant.datasets import download_datasets
 
 
 class PairwiseGrangerTestCase(unittest.TestCase):
@@ -453,12 +453,16 @@ def test_pairwise_spectral_granger_against_ground_truth(self):
             ("noise_covariance.npy", "6f80ccff2b2aa9485dc9c01d81570bf5")
         ]
 
+        downloaded_files = {}
         for filename, checksum in files_to_download:
-            download_datasets(repo_path=f"{repo_path}/{filename}",
-                              checksum=checksum)
-        signals = np.load(ELEPHANT_TMP_DIR / 'time_series.npy')
-        weights = np.load(ELEPHANT_TMP_DIR / 'weights.npy')
-        cov = np.load(ELEPHANT_TMP_DIR / 'noise_covariance.npy')
+            downloaded_files[filename] = {
+                'filename': filename,
+                'path': download_datasets(repo_path=f"{repo_path}/{filename}",
+                                          checksum=checksum)}
+
+        signals = np.load(downloaded_files['time_series.npy']['path'])
+        weights = np.load(downloaded_files['weights.npy']['path'])
+        cov = np.load(downloaded_files['noise_covariance.npy']['path'])
 
         # Estimate spectral Granger Causality
         f, spectral_causality = \
@@ -532,11 +536,15 @@ def test_pairwise_spectral_granger_against_r_grangers(self):
             ("gc_matrix.npy", "c57262145e74a178588ff0a1004879e2")
         ]
 
+        downloaded_files = {}
         for filename, checksum in files_to_download:
-            download_datasets(repo_path=f"{repo_path}/{filename}",
-                              checksum=checksum)
-        signal = np.load(ELEPHANT_TMP_DIR / 'time_series_small.npy')
-        gc_matrix = np.load(ELEPHANT_TMP_DIR / 'gc_matrix.npy')
+            downloaded_files[filename] = {
+                'filename': filename,
+                'path': download_datasets(repo_path=f"{repo_path}/{filename}",
+                                          checksum=checksum)}
+
+        signal = np.load(downloaded_files['time_series_small.npy']['path'])
+        gc_matrix = np.load(downloaded_files['gc_matrix.npy']['path'])
 
         denom = 20
         f, spectral_causality = \
diff --git a/elephant/test/test_datasets.py b/elephant/test/test_datasets.py
new file mode 100644
index 000000000..24334e2f6
--- /dev/null
+++ b/elephant/test/test_datasets.py
@@ -0,0 +1,35 @@
+import unittest
+import os
+from unittest.mock import patch
+from pathlib import Path
+import urllib
+
+from elephant.datasets import download_datasets
+
+
+class TestDownloadDatasets(unittest.TestCase):
+    @patch.dict(os.environ, {'ELEPHANT_DATA_LOCATION': '/valid/path'}, clear=True)
+    @patch('os.path.exists', return_value=True)
+    def test_valid_path(self, mock_exists):
+        repo_path = 'some/repo/path'
+        expected = Path('/valid/path/some/repo/path')
+        result = download_datasets(repo_path)
+        self.assertEqual(result, expected)
+
+    @patch.dict(os.environ, {'ELEPHANT_DATA_LOCATION': 'http://valid.url'}, clear=True)
+    @patch('os.path.exists', return_value=False)
+    def test_valid_url(self, mock_exists):
+        repo_path = 'some/repo/path'
+        self.assertRaises(urllib.error.URLError, download_datasets, repo_path)
+
+    @patch.dict(os.environ, {'ELEPHANT_DATA_LOCATION': 'invalid_path_or_url'}, clear=True)
+    @patch('os.path.exists', return_value=False)
+    def test_invalid_value(self, mock_exists):
+        repo_path = 'some/repo/path'
+        with self.assertRaises(ValueError) as cm:
+            download_datasets(repo_path)
+        self.assertIn("invalid_path_or_url", str(cm.exception))
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/elephant/test/test_phase_analysis.py b/elephant/test/test_phase_analysis.py
index 6b8218c99..c328681ec 100644
--- a/elephant/test/test_phase_analysis.py
+++ b/elephant/test/test_phase_analysis.py
@@ -384,6 +384,7 @@ def setUpClass(cls):
         # downloaded once into a local temporary directory
         # and then loaded/ read for each test function individually.
 
+        cls.tmp_path = {}
         # REAL DATA
         real_data_path = "unittest/phase_analysis/weighted_phase_lag_index/" \
                          "data/wpli_real_data"
@@ -398,8 +399,10 @@ def setUpClass(cls):
         )
         for filename, checksum in cls.files_to_download_real:
             # files will be downloaded to ELEPHANT_TMP_DIR
-            cls.tmp_path = download_datasets(
-                f"{real_data_path}/{filename}", checksum=checksum)
+            cls.tmp_path[filename] = {
+                'filename': filename,
+                'path': download_datasets(
+                    f"{real_data_path}/{filename}", checksum=checksum)}
         # ARTIFICIAL DATA
         artificial_data_path = "unittest/phase_analysis/" \
             "weighted_phase_lag_index/data/wpli_specific_artificial_dataset"
@@ -409,11 +412,13 @@ def setUpClass(cls):
         )
         for filename, checksum in cls.files_to_download_artificial:
             # files will be downloaded to ELEPHANT_TMP_DIR
-            cls.tmp_path = download_datasets(
-                f"{artificial_data_path}/{filename}", checksum=checksum)
+            cls.tmp_path[filename] = {
+                'filename': filename,
+                'path': download_datasets(
+                    f"{artificial_data_path}/{filename}", checksum=checksum)}
         # GROUND TRUTH DATA
         ground_truth_data_path = "unittest/phase_analysis/" \
-                        "weighted_phase_lag_index/data/wpli_ground_truth"
+            "weighted_phase_lag_index/data/wpli_ground_truth"
         cls.files_to_download_ground_truth = (
             ("ground_truth_WPLI_from_ft_connectivity_wpli_"
              "with_real_LFPs_R2G.csv", "4d9a7b7afab7d107023956077ab11fef"),
@@ -422,8 +427,10 @@ def setUpClass(cls):
         )
         for filename, checksum in cls.files_to_download_ground_truth:
             # files will be downloaded into ELEPHANT_TMP_DIR
-            cls.tmp_path = download_datasets(
-                f"{ground_truth_data_path}/{filename}", checksum=checksum)
+            cls.tmp_path[filename] = {
+                'filename': filename,
+                'path': download_datasets(
+                    f"{ground_truth_data_path}/{filename}", checksum=checksum)}
 
     def setUp(self):
         self.tolerance = 1e-15
@@ -431,10 +438,10 @@ def setUp(self):
         # load real/artificial LFP-dataset for ground-truth consistency checks
         # real LFP-dataset
         dataset1_real = scipy.io.loadmat(
-            f"{self.tmp_path.parent}/{self.files_to_download_real[0][0]}",
+            f"{self.tmp_path[self.files_to_download_real[0][0]]['path']}",
             squeeze_me=True)
         dataset2_real = scipy.io.loadmat(
-            f"{self.tmp_path.parent}/{self.files_to_download_real[1][0]}",
+            f"{self.tmp_path[self.files_to_download_real[1][0]]['path']}",
             squeeze_me=True)
 
         # get relevant values
@@ -449,12 +456,12 @@ def setUp(self):
             signal=self.lfps2_real, sampling_rate=self.sf2_real)
 
         # artificial LFP-dataset
-        dataset1_artificial = scipy.io.loadmat(
-            f"{self.tmp_path.parent}/"
-            f"{self.files_to_download_artificial[0][0]}", squeeze_me=True)
-        dataset2_artificial = scipy.io.loadmat(
-            f"{self.tmp_path.parent}/"
-            f"{self.files_to_download_artificial[1][0]}", squeeze_me=True)
+        dataset1_path = \
+            f"{self.tmp_path[self.files_to_download_artificial[0][0]]['path']}"
+        dataset1_artificial = scipy.io.loadmat(dataset1_path, squeeze_me=True)
+        dataset2_path = \
+            f"{self.tmp_path[self.files_to_download_artificial[1][0]]['path']}"
+        dataset2_artificial = scipy.io.loadmat(dataset2_path, squeeze_me=True)
         # get relevant values
         self.lfps1_artificial = dataset1_artificial['lfp_matrix'] * pq.uV
         self.sf1_artificial = dataset1_artificial['sf'] * pq.Hz
@@ -469,12 +476,10 @@ def setUp(self):
         # load ground-truth reference calculated by:
         # Matlab package 'FieldTrip': ft_connectivity_wpli()
         self.wpli_ground_truth_ft_connectivity_wpli_real = np.loadtxt(
-            f"{self.tmp_path.parent}/"
-            f"{self.files_to_download_ground_truth[0][0]}",
+            f"{self.tmp_path[self.files_to_download_ground_truth[0][0]]['path']}",  # noqa
             delimiter=',', dtype=np.float64)
         self.wpli_ground_truth_ft_connectivity_artificial = np.loadtxt(
-            f"{self.tmp_path.parent}/"
-            f"{self.files_to_download_ground_truth[1][0]}",
+            f"{self.tmp_path[self.files_to_download_ground_truth[1][0]]['path']}",  # noqa
             delimiter=',', dtype=np.float64)
 
     def test_WPLI_ground_truth_consistency_real_LFP_dataset(self):
@@ -700,7 +705,7 @@ def test_WPLI_raises_error_if_AnalogSignals_have_diff_sampling_rate():
     def test_WPLI_raises_error_if_sampling_rate_not_given(self):
         """
         Test if WPLI raises a ValueError, when the sampling rate is not given
-        for np.array() or Quanitity input.
+        for np.array() or Quantity input.
         """
         signal_x = np.random.random([40, 2100]) * pq.mV
         signal_y = np.random.random([40, 2100]) * pq.mV
diff --git a/elephant/test/test_spectral.py b/elephant/test/test_spectral.py
index 41244fb66..00cdd6941 100644
--- a/elephant/test/test_spectral.py
+++ b/elephant/test/test_spectral.py
@@ -20,7 +20,7 @@
 from packaging import version
 
 import elephant.spectral
-from elephant.datasets import download_datasets, ELEPHANT_TMP_DIR
+from elephant.datasets import download_datasets
 
 
 class WelchPSDTestCase(unittest.TestCase):
@@ -278,12 +278,15 @@ def test_multitaper_psd_against_nitime(self):
             ("psd_nitime.npy", "89d1f53957e66c786049ea425b53c0e8")
         ]
 
+        downloaded_files = {}
         for filename, checksum in files_to_download:
-            download_datasets(repo_path=f"{repo_path}/{filename}",
-                              checksum=checksum)
+            downloaded_files[filename] = {
+                'filename': filename,
+                'path': download_datasets(repo_path=f"{repo_path}/{filename}",
+                                          checksum=checksum)}
 
-        time_series = np.load(ELEPHANT_TMP_DIR / 'time_series.npy')
-        psd_nitime = np.load(ELEPHANT_TMP_DIR / 'psd_nitime.npy')
+        time_series = np.load(downloaded_files['time_series.npy']['path'])
+        psd_nitime = np.load(downloaded_files['psd_nitime.npy']['path'])
 
         freqs, psd_multitaper = elephant.spectral.multitaper_psd(
             signal=time_series, fs=0.1, nw=4, num_tapers=8)
diff --git a/elephant/test/test_spike_train_correlation.py b/elephant/test/test_spike_train_correlation.py
index cae01e479..90de65ea1 100644
--- a/elephant/test/test_spike_train_correlation.py
+++ b/elephant/test/test_spike_train_correlation.py
@@ -20,7 +20,7 @@
 from elephant.spike_train_generation import StationaryPoissonProcess, \
     StationaryGammaProcess
 import math
-from elephant.datasets import download_datasets, ELEPHANT_TMP_DIR
+from elephant.datasets import download_datasets
 from elephant.spike_train_generation import homogeneous_poisson_process, \
     homogeneous_gamma_process
 
@@ -852,11 +852,15 @@ def test_sttc_validation_test(self):
         files_to_download = [("spike_time_tiling_coefficient_results.nix",
                               "e3749d79046622494660a03e89950f51")]
 
+        downloaded_files = {}
         for filename, checksum in files_to_download:
-            filepath = download_datasets(repo_path=f"{repo_path}/{filename}",
-                                         checksum=checksum)
+            downloaded_files[filename] = {
+                'filename': filename,
+                'path': download_datasets(repo_path=f"{repo_path}/{filename}",
+                                          checksum=checksum)}
 
-        reader = NixIO(filepath, mode='ro')
+        reader = NixIO(downloaded_files[
+            'spike_time_tiling_coefficient_results.nix']['path'], mode='ro')
         test_data_block = reader.read()
 
         for segment in test_data_block[0].segments:
diff --git a/elephant/test/test_spike_train_dissimilarity.py b/elephant/test/test_spike_train_dissimilarity.py
index 6cab2858b..4619d4bba 100644
--- a/elephant/test/test_spike_train_dissimilarity.py
+++ b/elephant/test/test_spike_train_dissimilarity.py
@@ -15,7 +15,7 @@
 from elephant.spike_train_generation import StationaryPoissonProcess
 import elephant.spike_train_dissimilarity as stds
 
-from elephant.datasets import download_datasets, ELEPHANT_TMP_DIR
+from elephant.datasets import download_datasets
 
 
 class TimeScaleDependSpikeTrainDissimMeasuresTestCase(unittest.TestCase):
@@ -398,12 +398,16 @@ def test_victor_purpura_matlab_comparison_float(self):
             ("times_float.npy", "ed1ff4d2c0eeed4a2b50a456803656be"),
             ("matlab_results_float.npy", "a17f049e7ad0ddf7ca812e86fdb92646")]
 
+        downloaded_files = {}
         for filename, checksum in files_to_download:
-            download_datasets(repo_path=f"{repo_path}/{filename}",
-                              checksum=checksum)
+            downloaded_files[filename] = {
+                'filename': filename,
+                'path': download_datasets(repo_path=f"{repo_path}/{filename}",
+                                          checksum=checksum)}
 
-        times_float = np.load(ELEPHANT_TMP_DIR / 'times_float.npy')
-        mat_res_float = np.load(ELEPHANT_TMP_DIR / 'matlab_results_float.npy')
+        times_float = np.load(downloaded_files['times_float.npy']['path'])
+        mat_res_float = np.load(downloaded_files[
+            'matlab_results_float.npy']['path'])
 
         r_float = SpikeTrain(times_float[0], units='ms', t_start=0,
                              t_stop=1000 * ms)
@@ -428,12 +432,16 @@ def test_victor_purpura_matlab_comparison_int(self):
             ("times_int.npy", "aa1411c04da3f58d8b8913ae2f935057"),
             ("matlab_results_int.npy", "7edd32e50edde12dc1ef4aa5f57f70fb")]
 
+        downloaded_files = {}
         for filename, checksum in files_to_download:
-            download_datasets(repo_path=f"{repo_path}/{filename}",
-                              checksum=checksum)
-
-        times_int = np.load(ELEPHANT_TMP_DIR / 'times_int.npy')
-        mat_res_int = np.load(ELEPHANT_TMP_DIR / 'matlab_results_int.npy')
+            downloaded_files[filename] = {
+                'filename': filename,
+                'path': download_datasets(repo_path=f"{repo_path}/{filename}",
+                                          checksum=checksum)}
+
+        times_int = np.load(downloaded_files['times_int.npy']['path'])
+        mat_res_int = np.load(
+            downloaded_files['matlab_results_int.npy']['path'])
 
         r_int = SpikeTrain(times_int[0], units='ms', t_start=0,
                            t_stop=1000 * ms)
diff --git a/elephant/test/test_spike_train_synchrony.py b/elephant/test/test_spike_train_synchrony.py
index 58be525eb..d3455b115 100644
--- a/elephant/test/test_spike_train_synchrony.py
+++ b/elephant/test/test_spike_train_synchrony.py
@@ -168,7 +168,7 @@ def test_spike_contrast_with_Izhikevich_network_auto(self):
         checksum = "70e848500c1d9c6403b66de8c741d849"
         filepath_zip = download_datasets(repo_path=izhikevich_gin,
                                          checksum=checksum)
-        unzip(filepath_zip)
+        unzip(filepath_zip, outdir=filepath_zip.parent)
         filepath_json = filepath_zip.with_suffix(".json")
         with open(filepath_json) as read_file:
             data = json.load(read_file)