diff --git a/tablib/core.py b/tablib/core.py
index 1bdd49f..80f4418 100644
--- a/tablib/core.py
+++ b/tablib/core.py
@@ -223,6 +223,23 @@ class Dataset(object):
         except AttributeError:
             return '<dataset object>'
 
+    def __unicode__(self):
+        result = [self.__headers]
+
+        result.extend(map(unicode, row) for row in self._data)
+
+        # here, we calculate max width for each column
+        lens = (map(len, row) for row in result)
+        field_lens = map(max, zip(*lens))
+
+        # delimiter between header and data
+        result.insert(1, [u'-' * length for length in field_lens])
+
+        format_string = u'|'.join(u'{%s:%s}' % item for item in enumerate(field_lens))
+
+        return u'\n'.join(format_string.format(*row) for row in result)
+
+
 
     # ---------
     # Internals
diff --git a/test_tablib.py b/test_tablib.py
index 5455c28..0bfb417 100755
--- a/test_tablib.py
+++ b/test_tablib.py
@@ -649,6 +649,22 @@ class TablibTestCase(unittest.TestCase):
 
         self.assertEquals(self.founders[orig_target_header], data[target_header])
 
+    def test_unicode_renders_markdown_table(self):
+        # add another entry to test right field width for
+        # integer
+        self.founders.append(('Old', 'Man', 100500))
+
+        self.assertEquals(
+            u"""
+first_name|last_name |gpa   
+----------|----------|------
+John      |Adams     |90    
+George    |Washington|67    
+Thomas    |Jefferson |50    
+Old       |Man       |100500
+""".strip(),
+            unicode(self.founders)
+        )
 
 
 if __name__ == '__main__':