diff --git a/src/cristallina/analysis.py b/src/cristallina/analysis.py
new file mode 100644
index 0000000..449bd6c
--- /dev/null
+++ b/src/cristallina/analysis.py
@@ -0,0 +1,64 @@
+import re
+from collections import defaultdict
+
+import numpy as np
+
+from sfdata import SFDataFiles, sfdatafile, SFScanInfo
+
+import joblib
+from joblib import Parallel, delayed, Memory
+
+# TODO: generalize this for all analysis, i.e. find appropriate p-group
+location = "/das/work/units/cristallina/p19739/cachedir"
+memory = Memory(location, verbose=0, compress=2)
+
+import jungfrau_utils as ju
+ju_patch_less_verbose(ju)
+
+from .utils import ROI
+
+
+@memory.cache
+def sum_images(fileset, channel="JF16T03V01", alignment_channels=None, batch_size=10, roi : ROI = None, preview=False):
+    """
+    Sums a given region of interest (roi) for an image channel from a
+    given fileset (e.g. "run0352/data/acq0001.*.h5").
+
+    Allows alignment, i.e. reducing only to a common subset with other channels.
+
+    Summation is performed in batches, preview only sums and returns the first batch.
+    """
+
+    with SFDataFiles(*fileset) as data:
+        if alignment_channels is not None:
+            channels = [channel] + [ch for ch in alignment_channels]
+        else:
+            channels = [channel]
+
+        subset = data[channels]
+
+        subset.drop_missing()
+
+        Images = subset[channel]
+
+        res = defaultdict(list)
+        res["roi"] = repr(roi)
+
+        for image_slice in Images.in_batches(batch_size):
+
+            index_slice, im = image_slice
+
+            if roi is None:
+                im_ROI = im[:]
+            else:
+                im_ROI = im[:, roi.rows, roi.cols]
+
+            res[f"{channel}_intensity"].append(np.sum(im_ROI, axis=(1, 2)))
+            res["pids"].append(Images.pids[index_slice])
+
+            # only return first batch
+            if preview:
+                break
+
+    return res
+
diff --git a/src/cristallina/plot.py b/src/cristallina/plot.py
new file mode 100644
index 0000000..915a7b1
--- /dev/null
+++ b/src/cristallina/plot.py
@@ -0,0 +1,190 @@
+import re
+from collections import defaultdict
+
+import matplotlib
+from matplotlib import pyplot as plt
+
+import warnings
+# because of https://github.com/kornia/kornia/issues/1425
+warnings.simplefilter("ignore", DeprecationWarning)
+
+import numpy as np
+from tqdm import tqdm
+from matplotlib import patches
+
+from pathlib import Path
+
+from sfdata import SFDataFiles, sfdatafile, SFScanInfo
+import jungfrau_utils as ju
+
+from .utils import ROI
+
+def ju_patch_less_verbose(ju_module):
+    """Quick monkey patch to suppress verbose messages from gain & pedestal file searcher."""
+    ju_module.swissfel_helpers._locate_gain_file = ju_module.swissfel_helpers.locate_gain_file
+    ju_module.swissfel_helpers._locate_pedestal_file = ju_module.swissfel_helpers.locate_pedestal_file
+
+    def less_verbose_gain(*args, **kwargs):
+        kwargs["verbose"] = False
+        return ju_module.swissfel_helpers._locate_gain_file(*args, **kwargs)
+
+    def less_verbose_pedestal(*args, **kwargs):
+        kwargs["verbose"] = False
+        return ju_module.swissfel_helpers._locate_pedestal_file(*args, **kwargs)
+
+    # ju_module.swissfel_helpers.locate_gain_file = less_verbose_gain
+    # ju_module.swissfel_helpers.locate_pedestal_file = less_verbose_pedestal
+
+    ju_module.file_adapter.locate_gain_file = less_verbose_gain
+    ju_module.file_adapter.locate_pedestal_file = less_verbose_pedestal
+
+
+ju_patch_less_verbose(ju)
+
+
+
+
+
+def plot_correlation(x, y, ax=None, **ax_kwargs):
+    """
+    Plots the correlation of x and y in a normalized scatterplot.
+    If no axis is given a figure and axis are created.
+
+    Returns: The axis object and the correlation coefficient between
+             x and y.
+    """
+
+    xstd = np.std(x)
+    ystd = np.std(y)
+
+    xnorm = (x - np.mean(x)) / xstd
+    ynorm = (y - np.mean(y)) / ystd
+
+    n = len(ys)
+    
+    r = 1 / (n) * sum(xnorm * ynorm)
+
+    if ax is None:
+        fig, ax = plt.subplots()
+
+    if ax_kwargs is not None:
+        ax.set(**ax_kwargs)
+
+    ax.plot(xnorm, ynorm, "o")
+    ax.text(0.95, 0.05, f"r = {r:.2f}", transform=ax.transAxes, horizontalalignment="right")
+
+    return ax, r
+
+def plot_channel(data, channel_name, ax=None):
+
+    channel_dim = len(data[channel_name].shape)
+    # dim == 3: a 2D Image
+    # dim == 2: an array per pulse (probably)
+    # dim == 1: a single value per pulse (probably)
+
+    plot_f = {
+        1: plot_1d_channel,
+        2: plot_2d_channel,
+        3: plot_image_channel,
+    }
+
+    plot_f[channel_dim](data, channel_name, ax)
+
+
+def axis_styling(ax, channel_name, description):
+
+    ax.set_title(channel_name)
+    # ax.set_xlabel('x')
+    # ax.set_ylabel('a.u.')
+    ax.ticklabel_format(useOffset=False)
+    ax.text(
+        0.05,
+        0.05,
+        description,
+        transform=ax.transAxes,
+        horizontalalignment="left",
+        bbox=dict(boxstyle="round", color="lightgrey"),
+    )
+
+
+def plot_1d_channel(data, channel_name, ax=None):
+    """
+    Plots channel data for a channel that contains a single numeric value per pulse.
+    """
+    try:
+        mean, std = np.mean(data[channel_name].data), np.std(data[channel_name].data)
+        n_entries_per_frame = data[channel_name].shape
+    except TypeError:
+        print(f"Cannot parse channel {channel_name}. Check dimensionality.")
+        return
+
+    y_data = data[channel_name].data
+
+    if ax is None:
+        fig, ax = plt.subplots(constrained_layout=True)
+
+    ax.plot(y_data)
+    description = f"mean: {mean:.2e},\nstd: {std:.2e}"
+    axis_styling(ax, channel_name, description)
+
+
+def plot_2d_channel(data, channel_name, ax=None):
+    """
+    Plots channel data for a channel that contains a 1d array of numeric values per pulse.
+    """
+    try:
+        mean, std = np.mean(data[channel_name].data), np.std(data[channel_name].data)
+        # data[channel_name].data
+        mean_over_frames = np.mean(data[channel_name].data, axis=0)
+    except TypeError:
+        print(f"Unknown data in channel {channel_name}.")
+        return
+
+    y_data = mean_over_frames
+
+    if ax is None:
+        fig, ax = plt.subplots(constrained_layout=True)
+
+    ax.plot(y_data)
+    description = f"mean: {mean:.2e},\nstd: {std:.2e}"
+    axis_styling(ax, channel_name, description)
+
+
+def plot_image_channel(data, channel_name, pulse=0, ax=None, rois=None, norms=None):
+    """
+    Plots channel data for a channel that contains an image (2d array) of numeric values per pulse.
+    """
+
+    im = data[channel_name].data[pulse]
+
+    if ax is None:
+        fig, ax = plt.subplots(constrained_layout=True)
+
+    std = im.std()
+    mean = im.mean()
+
+    if norms is None:
+        norm = matplotlib.colors.Normalize(vmin=mean - std, vmax=mean + std)
+    else:
+        norm = matplotlib.colors.Normalize(vmin=norms[0], vmax=norms[1])
+
+    ax.imshow(im, norm=norm)
+    ax.invert_yaxis()
+
+    if rois is not None:
+        # Plot rois if given
+        for i, roi in enumerate(rois):
+            # Create a rectangle with ([bottom left corner coordinates], width, height)
+            rect = patches.Rectangle(
+                [roi.left, roi.bottom], roi.width, roi.height,
+                linewidth=3,
+                edgecolor=f"C{i}",
+                facecolor="none",
+                label=roi.name,
+            )
+            ax.add_patch(rect)
+
+    description = f"mean: {mean:.2e},\nstd: {std:.2e}"
+    axis_styling(ax, channel_name, description)
+    plt.legend(loc=4)    
+
diff --git a/src/cristallina/utils.py b/src/cristallina/utils.py
new file mode 100644
index 0000000..03dbfc8
--- /dev/null
+++ b/src/cristallina/utils.py
@@ -0,0 +1,108 @@
+import yaml
+import os
+
+from sfdata import SFDataFiles, sfdatafile, SFScanInfo
+
+
+def print_run_info(
+    run_number=42, print_channels=True, extra_verbose=False, base_path="/sf/cristallina/data/p19739/raw/"
+):
+    """Prints overview of run information.
+
+    Extra verbose output contains all files and pids.
+    """
+
+    scan = SFScanInfo(f"{base_path}/run{run_number:04}/meta/scan.json")
+
+    short = {}
+    for key, value in scan.info.items():
+        if isinstance(value, list):
+            short[key] = value[:2]
+            short[key].append("...")
+            short[key].append(value[-1])
+        else:
+            short[key] = value
+
+    if extra_verbose:
+        print(yaml.dump(scan.info, sort_keys=False, default_flow_style=False))
+    else:
+        print(yaml.dump(short, sort_keys=False, default_flow_style=False))
+
+    print(f"Number of steps: {len(scan.info['scan_readbacks'])}")
+
+    total_size = 0
+    for files_in_step in scan.info["scan_files"]:
+        for file in files_in_step:
+            try:
+                total_size += os.path.getsize(file)
+            except FileNotFoundError as e:
+                pass
+
+    print(f"Total file size: {total_size/(1024*1024*1024):.1f} GB\n")
+
+    for step in scan:
+        ch = step.channels
+        print("\n".join([str(c) for c in ch]))
+        # print only channels for first step
+        break
+
+
+
+class ROI:
+    """Definition of region of interest (ROI) in image coordinates.
+
+    Example: ROI(left=10, right=20, bottom=100, top=200)
+    """
+
+    def __init__(
+        self,
+        left: int = None,
+        right: int = None,
+        top: int = None,
+        bottom: int = None,
+        center_x: int = None,
+        center_y: int = None,
+        width: int = None,
+        height: int = None,
+        name: str = None,
+    ):
+
+        if None not in (left, right, bottom, top):
+            self.left, self.right, self.bottom, self.top, = (
+                left,
+                right,
+                bottom,
+                top,
+            )
+        elif None not in (center_x, center_y, width, height):
+            self.from_centers_widths(center_x, center_y, width, height)
+        else:
+            raise ValueError("No valid ROI definition.")
+        
+        self.name = name
+
+    def from_centers_widths(self, center_x, center_y, width, height):
+        self.left = center_x - width // 2
+        self.right = center_x + width // 2
+
+        self.top = center_y + height // 2
+        self.bottom = center_y - height // 2
+
+    @property
+    def rows(self):
+        return slice(self.bottom, self.top)
+
+    @property
+    def cols(self):
+        return slice(self.left, self.right)
+    
+    @property
+    def width(self):
+        return self.right - self.left
+    
+    @property
+    def height(self):
+        return self.top - self.bottom
+    
+    def __repr__(self):
+        return f"ROI(bottom={self.bottom}, top={self.top}, left={self.left}, right={self.right})"
diff --git a/tests/conftest.py b/tests/conftest.py
index a4b4bda..5e41c6d 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -1,10 +1,29 @@
 """
-    Dummy conftest.py for cristallina.
+    conftest.py for cristallina.
 
-    If you don't know what this is for, just leave it empty.
     Read more about conftest.py under:
     - https://docs.pytest.org/en/stable/fixture.html
     - https://docs.pytest.org/en/stable/writing_plugins.html
 """
 
-# import pytest
+import pytest
+
+
+def pytest_addoption(parser):
+    parser.addoption(
+        "--runregression", action="store_true", default=False, help="run slow regression tests"
+    )
+
+
+def pytest_configure(config):
+    config.addinivalue_line("markers", "regression: mark test as a slow regression test to run")
+
+
+def pytest_collection_modifyitems(config, items):
+    if config.getoption("--runregression"):
+        # --runregression given in cli: do not skip slow tests
+        return
+    skip_regression = pytest.mark.skip(reason="need --runregression option to run")
+    for item in items:
+        if "regression" in item.keywords:
+            item.add_marker(skip_regression)
diff --git a/tests/test_regression.py b/tests/test_regression.py
new file mode 100644
index 0000000..99c9787
--- /dev/null
+++ b/tests/test_regression.py
@@ -0,0 +1,28 @@
+"""
+Regression tests for functionality that we assume to be working for our analysis. 
+"""
+
+import pytest
+from sfdata import SFDataFiles, sfdatafile, SFScanInfo
+import tracemalloc
+
+
+@pytest.mark.regression
+def test_JU_memory():
+    base_path = "/sf/cristallina/data/p19739/raw/"
+    run_number = 49
+    averages = []
+
+    tracemalloc.start()
+    current, peak = tracemalloc.get_traced_memory()
+    print(f"Current memory usage is {current / 10**6:.1f}MB; Peak was {peak / 10**6:.1f}MB")
+
+
+    with SFDataFiles(f"{base_path}/run{run_number:04}/data/acq00*.h5") as data:
+        ch = data["JF16T03V01"]
+            
+    current, peak = tracemalloc.get_traced_memory()
+    print(f"Current memory usage is {current / 10**6:.1f}MB; Peak was {peak / 10**6:.1f}MB")
+    tracemalloc.stop()   
+
+    assert current/10**6 < 100, "Memory consumption should be below 100 MB"
diff --git a/tests/test_utils.py b/tests/test_utils.py
new file mode 100644
index 0000000..0d5f020
--- /dev/null
+++ b/tests/test_utils.py
@@ -0,0 +1,21 @@
+import pytest
+
+from cristallina.utils import ROI, print_run_info
+
+__author__ = "Alexander Steppke"
+
+def test_print(capsys):
+
+    print_run_info(247)
+    captured = capsys.readouterr()
+    assert "15453208940" in captured.out
+    assert "LiTbF4_rocking" in captured.out
+
+
+	
+def test_ROI():
+    """API Tests"""
+    r = ROI(left=1, right=2, top=4, bottom=2)
+
+    assert r.width == 1
+    assert r.height == 2