diff --git a/projects/project_vector_strength/export_data.py b/projects/project_vector_strength/export_data.py
new file mode 100644
index 0000000..b3d465c
--- /dev/null
+++ b/projects/project_vector_strength/export_data.py
@@ -0,0 +1,115 @@
+import numpy as np
+import scipy.io as sp
+import subprocess
+import glob
+import sys, os
+sys.path.append(os.path.expanduser('~/code/python/pyRELACS'))
+from pyRELACS.DataLoader import iload
+from IPython import embed;
+
+def read_info_file(file_name):
+    """
+    Reads the info file and returns the stored metadata in a dictionary. The dictionary may be nested.
+    @param file_name:  The name of the info file.
+    @return: dictionary, the stored information.
+    """
+    information = []
+    root = {}
+    with open(file_name, 'r') as f:
+        lines = f.readlines()
+        for l in lines:
+            if not l.startswith("#"):
+                continue
+            l = l.strip("#").strip()
+            if len(l) == 0:
+                continue
+            if not ": " in l:
+                sec = {}
+                root[l] = sec
+            else:
+                parts = l.split(': ')
+                sec[parts[0].strip()] = parts[1].strip()
+    information.append(root)
+    return information
+
+
+def get_zero_crossings(trace, threshold=0.0, rising_flank=True):
+    if rising_flank:
+        shifted_trace = np.hstack((0, np.squeeze(trace[0:-1])))
+        crossings = np.all(np.vstack((np.squeeze(trace) > threshold, shifted_trace <= threshold)), 0)
+    else:
+        shifted_trace = np.hstack((0, np.squeeze(trace[0:-1])))
+        crossings = np.all(np.vstack((np.squeeze(trace) < threshold, shifted_trace >= threshold)), 0)
+    positions = np.nonzero(crossings)[0]
+    return positions
+
+
+def has_cell_type(cell, cell_type):
+    if not os.path.exists(cell + '/info.dat'):
+        return false
+    else:
+        info = read_info_file(cell + '/info.dat')
+        return info[0]['Cell']['CellType'].lower() == cell_type.lower()
+
+
+def has_baseline(cell):
+    return os.path.exists(cell + '/basespikes1.dat')
+
+
+def has_raw_traces(cell):
+    if os.path.exists(cell + '/trace-2.raw'):
+        return True
+    if os.path.exists(cell + '/trace-2.raw.gz'):
+        if subprocess.check_call(['gzip', '-d' , cell+'/trace-2.raw']) == 0:
+            return True
+    return False
+
+
+def load_eod_data(cell, max_time, sample_rate):
+    eod = np.fromfile(cell + "/trace-2.raw", np.float32)
+    eod = eod[:max_time*sample_rate]
+    return eod
+
+
+def load_spike_data(cell, max_time):
+    data = list(iload(cell+'/basespikes1.dat'))
+    if data[0][2][-1]/1000 < max_time:
+        return None
+    else:
+        return data[0][2][data[0][2] <= max_time]/1000
+
+
+if __name__ == '__main__':
+    cell_type = 'p-unit'
+    directory = os.path.expanduser('~/data/apteronotus')
+    cells = glob.glob(directory + '/2012-*')
+    
+    found = 0
+    index = 0
+    max_cell_count = 25
+    while found < max_cell_count and index < len(cells):
+        cell = cells[index]
+        print cell
+        if not has_cell_type(cell, cell_type):
+            print 'wrong type' 
+            index += 1
+            continue
+        if not has_baseline(cell):
+            print 'no baseline'
+            index += 1
+            continue
+        if not has_raw_traces(cell):
+            print 'no raw trace'
+            index += 1
+            continue
+        spikes = load_spike_data(cell, 20)
+        if spikes is None:
+            print 'not enough baseline spikes'
+            index += 1
+            continue
+            
+        eod_data = load_eod_data(cell, 20, 200000)
+        name = cell.split('/')[-1] + '_baseline.mat'
+        sp.savemat(name, {'spike_times':spikes, 'eod':eod_data})
+        found += 1
+        index += 1