diff --git a/.vscode/settings.json b/.vscode/settings.json
index 6f25bfe..8486715 100644
--- a/.vscode/settings.json
+++ b/.vscode/settings.json
@@ -3,9 +3,5 @@
   "python.testing.unittestEnabled": false,
   "python.testing.pytestEnabled": true,
   "python.defaultInterpreterPath": "${workspaceFolder}/.venv/bin/python",
-  "python.formatting.blackPath": "${workspaceFolder}/.venv/bin/black",
   "python.terminal.activateEnvironment": true,
-  "python.linting.flake8Enabled": true,
-  "python.linting.pycodestyleEnabled": false,
-  "python.linting.enabled": true
 }
\ No newline at end of file
diff --git a/docs/api/data.md b/docs/api/data.md
new file mode 100644
index 0000000..631e5cb
--- /dev/null
+++ b/docs/api/data.md
@@ -0,0 +1,5 @@
+# Data classes and containers
+
+## `utilz.data`
+
+::: utilz.data
diff --git a/docs/api/generators.md b/docs/api/generators.md
new file mode 100644
index 0000000..8aecac7
--- /dev/null
+++ b/docs/api/generators.md
@@ -0,0 +1,5 @@
+# Generator tools
+
+## `utilz.genz`
+
+::: utilz.genz
diff --git a/docs/intro.ipynb b/docs/intro.ipynb
index 3e81f44..776016c 100644
--- a/docs/intro.ipynb
+++ b/docs/intro.ipynb
@@ -26,7 +26,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 16,
+   "execution_count": 1,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1420,7 +1420,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [
     {
@@ -1454,103 +1454,103 @@
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>0</th>\n",
-       "      <td>0.045894</td>\n",
-       "      <td>0.093716</td>\n",
-       "      <td>0.932221</td>\n",
+       "      <td>0.897455</td>\n",
+       "      <td>0.329248</td>\n",
+       "      <td>0.190562</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>1.310156</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
-       "      <td>0.738293</td>\n",
-       "      <td>0.249943</td>\n",
-       "      <td>0.518687</td>\n",
+       "      <td>0.411200</td>\n",
+       "      <td>0.151263</td>\n",
+       "      <td>0.204226</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-1.391970</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>2</th>\n",
-       "      <td>0.357182</td>\n",
-       "      <td>0.454217</td>\n",
-       "      <td>0.575472</td>\n",
+       "      <td>0.670361</td>\n",
+       "      <td>0.213199</td>\n",
+       "      <td>0.398662</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.048193</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>3</th>\n",
-       "      <td>0.289010</td>\n",
-       "      <td>0.453426</td>\n",
-       "      <td>0.211871</td>\n",
+       "      <td>0.590188</td>\n",
+       "      <td>0.940737</td>\n",
+       "      <td>0.826784</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-0.397329</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>4</th>\n",
-       "      <td>0.328628</td>\n",
-       "      <td>0.396641</td>\n",
-       "      <td>0.041587</td>\n",
+       "      <td>0.739239</td>\n",
+       "      <td>0.175956</td>\n",
+       "      <td>0.304016</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.430950</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>5</th>\n",
-       "      <td>0.481833</td>\n",
-       "      <td>0.394005</td>\n",
-       "      <td>0.503150</td>\n",
+       "      <td>0.708524</td>\n",
+       "      <td>0.960608</td>\n",
+       "      <td>0.286470</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.103200</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>6</th>\n",
-       "      <td>0.430750</td>\n",
-       "      <td>0.769627</td>\n",
-       "      <td>0.838887</td>\n",
+       "      <td>0.851708</td>\n",
+       "      <td>0.004294</td>\n",
+       "      <td>0.302206</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.635292</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>7</th>\n",
-       "      <td>0.882731</td>\n",
-       "      <td>0.122181</td>\n",
-       "      <td>0.393370</td>\n",
+       "      <td>0.309853</td>\n",
+       "      <td>0.954225</td>\n",
+       "      <td>0.954408</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-1.378318</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>8</th>\n",
-       "      <td>0.622302</td>\n",
-       "      <td>0.943480</td>\n",
-       "      <td>0.715790</td>\n",
+       "      <td>0.535253</td>\n",
+       "      <td>0.212095</td>\n",
+       "      <td>0.627933</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-0.540699</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>9</th>\n",
-       "      <td>0.419627</td>\n",
-       "      <td>0.882003</td>\n",
-       "      <td>0.629938</td>\n",
+       "      <td>0.998427</td>\n",
+       "      <td>0.934565</td>\n",
+       "      <td>0.602804</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>1.180524</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "text/plain": [
-       "         A1        B1        C1 group A1_normed_by_group\n",
-       "0  0.045894  0.093716  0.932221     A                 A1\n",
-       "1  0.738293  0.249943  0.518687     A                 A1\n",
-       "2  0.357182  0.454217  0.575472     A                 A1\n",
-       "3  0.289010  0.453426  0.211871     A                 A1\n",
-       "4  0.328628  0.396641  0.041587     A                 A1\n",
-       "5  0.481833  0.394005  0.503150     B                 A1\n",
-       "6  0.430750  0.769627  0.838887     B                 A1\n",
-       "7  0.882731  0.122181  0.393370     B                 A1\n",
-       "8  0.622302  0.943480  0.715790     B                 A1\n",
-       "9  0.419627  0.882003  0.629938     B                 A1"
+       "         A1        B1        C1 group  A1_normed_by_group\n",
+       "0  0.897455  0.329248  0.190562     A            1.310156\n",
+       "1  0.411200  0.151263  0.204226     A           -1.391970\n",
+       "2  0.670361  0.213199  0.398662     A            0.048193\n",
+       "3  0.590188  0.940737  0.826784     A           -0.397329\n",
+       "4  0.739239  0.175956  0.304016     A            0.430950\n",
+       "5  0.708524  0.960608  0.286470     B            0.103200\n",
+       "6  0.851708  0.004294  0.302206     B            0.635292\n",
+       "7  0.309853  0.954225  0.954408     B           -1.378318\n",
+       "8  0.535253  0.212095  0.627933     B           -0.540699\n",
+       "9  0.998427  0.934565  0.602804     B            1.180524"
       ]
      },
-     "execution_count": 5,
+     "execution_count": 2,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -1577,7 +1577,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [
     {
@@ -1612,113 +1612,125 @@
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>0</th>\n",
-       "      <td>0.045894</td>\n",
-       "      <td>0.093716</td>\n",
-       "      <td>0.932221</td>\n",
+       "      <td>0.897455</td>\n",
+       "      <td>0.329248</td>\n",
+       "      <td>0.190562</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>1.310156</td>\n",
+       "      <td>0.235766</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
-       "      <td>0.738293</td>\n",
-       "      <td>0.249943</td>\n",
-       "      <td>0.518687</td>\n",
+       "      <td>0.411200</td>\n",
+       "      <td>0.151263</td>\n",
+       "      <td>0.204226</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-1.391970</td>\n",
+       "      <td>-0.250488</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>2</th>\n",
-       "      <td>0.357182</td>\n",
-       "      <td>0.454217</td>\n",
-       "      <td>0.575472</td>\n",
+       "      <td>0.670361</td>\n",
+       "      <td>0.213199</td>\n",
+       "      <td>0.398662</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.048193</td>\n",
+       "      <td>0.008672</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>3</th>\n",
-       "      <td>0.289010</td>\n",
-       "      <td>0.453426</td>\n",
-       "      <td>0.211871</td>\n",
+       "      <td>0.590188</td>\n",
+       "      <td>0.940737</td>\n",
+       "      <td>0.826784</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-0.397329</td>\n",
+       "      <td>-0.071500</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>4</th>\n",
-       "      <td>0.328628</td>\n",
-       "      <td>0.396641</td>\n",
-       "      <td>0.041587</td>\n",
+       "      <td>0.739239</td>\n",
+       "      <td>0.175956</td>\n",
+       "      <td>0.304016</td>\n",
        "      <td>A</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.430950</td>\n",
+       "      <td>0.077551</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>5</th>\n",
-       "      <td>0.481833</td>\n",
-       "      <td>0.394005</td>\n",
-       "      <td>0.503150</td>\n",
+       "      <td>0.708524</td>\n",
+       "      <td>0.960608</td>\n",
+       "      <td>0.286470</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.103200</td>\n",
+       "      <td>0.027771</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>6</th>\n",
-       "      <td>0.430750</td>\n",
-       "      <td>0.769627</td>\n",
-       "      <td>0.838887</td>\n",
+       "      <td>0.851708</td>\n",
+       "      <td>0.004294</td>\n",
+       "      <td>0.302206</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>0.635292</td>\n",
+       "      <td>0.170955</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>7</th>\n",
-       "      <td>0.882731</td>\n",
-       "      <td>0.122181</td>\n",
-       "      <td>0.393370</td>\n",
+       "      <td>0.309853</td>\n",
+       "      <td>0.954225</td>\n",
+       "      <td>0.954408</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-1.378318</td>\n",
+       "      <td>-0.370900</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>8</th>\n",
-       "      <td>0.622302</td>\n",
-       "      <td>0.943480</td>\n",
-       "      <td>0.715790</td>\n",
+       "      <td>0.535253</td>\n",
+       "      <td>0.212095</td>\n",
+       "      <td>0.627933</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>-0.540699</td>\n",
+       "      <td>-0.145500</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>9</th>\n",
-       "      <td>0.419627</td>\n",
-       "      <td>0.882003</td>\n",
-       "      <td>0.629938</td>\n",
+       "      <td>0.998427</td>\n",
+       "      <td>0.934565</td>\n",
+       "      <td>0.602804</td>\n",
        "      <td>B</td>\n",
-       "      <td>A1</td>\n",
-       "      <td>A1</td>\n",
+       "      <td>1.180524</td>\n",
+       "      <td>0.317675</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "text/plain": [
-       "         A1        B1        C1 group A1_normed_by_group A1_centered_by_group\n",
-       "0  0.045894  0.093716  0.932221     A                 A1                   A1\n",
-       "1  0.738293  0.249943  0.518687     A                 A1                   A1\n",
-       "2  0.357182  0.454217  0.575472     A                 A1                   A1\n",
-       "3  0.289010  0.453426  0.211871     A                 A1                   A1\n",
-       "4  0.328628  0.396641  0.041587     A                 A1                   A1\n",
-       "5  0.481833  0.394005  0.503150     B                 A1                   A1\n",
-       "6  0.430750  0.769627  0.838887     B                 A1                   A1\n",
-       "7  0.882731  0.122181  0.393370     B                 A1                   A1\n",
-       "8  0.622302  0.943480  0.715790     B                 A1                   A1\n",
-       "9  0.419627  0.882003  0.629938     B                 A1                   A1"
+       "         A1        B1        C1 group  A1_normed_by_group  \\\n",
+       "0  0.897455  0.329248  0.190562     A            1.310156   \n",
+       "1  0.411200  0.151263  0.204226     A           -1.391970   \n",
+       "2  0.670361  0.213199  0.398662     A            0.048193   \n",
+       "3  0.590188  0.940737  0.826784     A           -0.397329   \n",
+       "4  0.739239  0.175956  0.304016     A            0.430950   \n",
+       "5  0.708524  0.960608  0.286470     B            0.103200   \n",
+       "6  0.851708  0.004294  0.302206     B            0.635292   \n",
+       "7  0.309853  0.954225  0.954408     B           -1.378318   \n",
+       "8  0.535253  0.212095  0.627933     B           -0.540699   \n",
+       "9  0.998427  0.934565  0.602804     B            1.180524   \n",
+       "\n",
+       "   A1_centered_by_group  \n",
+       "0              0.235766  \n",
+       "1             -0.250488  \n",
+       "2              0.008672  \n",
+       "3             -0.071500  \n",
+       "4              0.077551  \n",
+       "5              0.027771  \n",
+       "6              0.170955  \n",
+       "7             -0.370900  \n",
+       "8             -0.145500  \n",
+       "9              0.317675  "
       ]
      },
-     "execution_count": 6,
+     "execution_count": 3,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -1726,6 +1738,13 @@
    "source": [
     "new_df.norm_by_group('group', 'A1', scale=False)"
    ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {
diff --git a/mkdocs.yml b/mkdocs.yml
index 2cbed0e..e204349 100644
--- a/mkdocs.yml
+++ b/mkdocs.yml
@@ -25,12 +25,14 @@ nav:
       - I/O: api/io.md
       - Function ops: api/ops.md
       - Maps: api/maps.md
+      - Generators: api/generators.md
       - Shorthands: api/shorts.md
       - Pipes: api/pipes.md
       - Dataframe verbs and tools: api/df.md
       - Plotting tools: api/plot.md
       - Boilerplate helpers: api/boilerplate.md
       - Decorators: api/decorators.md
+      - Dataclasses: api/data.md
 plugins:
   - mkdocs-jupyter:
       ignore: ["dev/*"]
diff --git a/utilz/__init__.py b/utilz/__init__.py
index 85d7224..4f858f3 100644
--- a/utilz/__init__.py
+++ b/utilz/__init__.py
@@ -8,3 +8,5 @@
 from .boilerplate import *
 from .stats import *
 from .pipes import *
+from .data import *
+from .genz import *
diff --git a/utilz/data.py b/utilz/data.py
new file mode 100644
index 0000000..1d308d4
--- /dev/null
+++ b/utilz/data.py
@@ -0,0 +1,132 @@
+"""
+Data containers
+"""
+
+from .maps import map
+
+
+class Box(list):
+
+    """
+    Box is a flexible list-like container that allows for dot-notation access to attributes of its elements and methods. This makes it easy for example to perform a `.groupby()` operation on a list of dataframes.
+
+    Boxes can be transparent or opaque. Transparent boxes always return the result of an operation as a list. This is useful for example when you want to call a method on each box element and immediately work with the results.
+
+    Opaque boxes always return a new `Box` who's contents can be accessed using `.contents()` or slice notation `box[:]`. This allows for method chaining on the underlying data.
+
+    Examples:
+        >>> # Transparent box by default
+        >>> box = Box([df1, df2, df3])
+
+        >>> # Access content like a list
+        >>> box[0] # returns df1
+        >>> box[:] # returns [df1, df2, df3]
+        >>> box.contents() # equivalent
+
+        >>> # Access attributes or call methods just like you would on a single object
+        >>> box.head(10) # returns a list each df head
+        >>> box.shape # returns a list each df shape
+
+        >>> # Opaque box facilitates method chaining but need
+        >>> # `.contents()` to access results
+        >>> black_box = Box([df1, df2, df3], transparent=False)
+        >>> black_box.groupby('col').mean().contents()
+
+        >>> # Apply arbitrary functions to box elements
+        >>> result = box.map(lambda x: x + 1)
+
+        >>> # Can also modify in place without returning anything
+        >>> box.map(lambda x: x + 1, inplace=True)
+
+        >>> # Map respects box transparency for method chaining
+        >>> box.set_transparent(False)
+        >>> result = box.map(lambda x: x + 1).head().contents()
+    """
+
+    def __init__(self, iterable, transparent=True):
+        """
+        Create a new box from an iterable
+
+        Args:
+            list (iterable): iterable of objects to store in the box
+            transparent (bool): whether methods should return results (`True`) or a new box (`False`); Default True
+        """
+        super().__init__(iterable)
+        self._transparent_box = transparent
+
+    def __getattr__(self, name, *args, **kwargs):
+        if hasattr(self[0], name):
+            attr_or_method = getattr(self[0], name)
+            if callable(attr_or_method):
+
+                def fn(*args, **kwargs):
+                    out = []
+                    for elem in self:
+                        result = getattr(elem, name)
+                        result = result(*args, **kwargs) if callable(result) else result
+                        out.append(result)
+                    out = (
+                        out
+                        if self._transparent_box
+                        else Box(out, transparent=self._transparent_box)
+                    )
+                    return out
+
+                return fn
+            else:
+                out = [getattr(elem, name) for elem in self]
+                out = (
+                    out
+                    if self._transparent_box
+                    else Box(out, transparent=self._transparent_box)
+                )
+                return out
+
+        else:
+            raise AttributeError
+
+    def __repr__(self):
+        return f"Box(len={len(self)}, transparent={self._transparent_box}, type={self[0].__class__.__module__}.{self[0].__class__.__name__})"
+
+    def map(self, fn, inplace=False):
+        """
+        Apply a function to each element in the box
+
+        Args:
+            fn (callable): function to apply to each element
+            inplace (bool, optional): whether to modify the box in place or return a new box. Defaults to False.
+            *args: positional arguments to pass to `fn`
+            **kwargs: keyword arguments to pass to `fn`
+
+        Returns:
+            Box: new box with the results of applying `fn` to each element
+        """
+
+        out = map(fn, self)
+        if inplace:
+            self.__init__(out, transparent=self._transparent_box)
+        else:
+            out = (
+                out
+                if self._transparent_box
+                else Box(out, transparent=self._transparent_box)
+            )
+            return out
+
+    def contents(self):
+        """
+        Convert box to list
+
+        Returns:
+            list: list of elements
+        """
+        return list(self)
+
+    def set_transparent(self, transparent):
+        """
+        Set the transparency of the box
+
+        Args:
+            transparent (bool): whether the box should be transparent or not
+        """
+        self._transparent_box = transparent
diff --git a/utilz/dev/data.py b/utilz/dev/data.py
new file mode 100644
index 0000000..a41305f
--- /dev/null
+++ b/utilz/dev/data.py
@@ -0,0 +1,115 @@
+# %%
+import numpy as np
+import pandas as pd
+from pathlib import Path
+from dataclasses import dataclass
+from functools import cached_property
+from utilz import randdf, equal
+from shutil import rmtree
+from utilz import discard
+
+
+# %% Subject data container
+@dataclass
+class Subject:
+    """Data container that makes uses of new cached_property decorator"""
+
+    sid: str
+    data_dir: Path
+
+    def __post_init__(self):
+        self.data_dir = self.data_dir / self.sid
+
+    @cached_property
+    def data(self):
+        print("Loading data")
+        return pd.read_csv(self.data_dir / "data.csv")
+
+    def clear_cache(self):
+        del self.data
+
+
+# %% Subject data test
+def test_subject():
+    # Setup data folders
+    DATA_DIR = Path(".")
+    SID = "sid001"
+    SUB_PATH = DATA_DIR / SID
+    SUB_PATH.mkdir(exist_ok=True)
+    FNAME = "data.csv"
+
+    # Make data
+    data = randdf()
+    data.to_csv(SUB_PATH / FNAME, index=False)
+
+    # Create subject
+    s = Subject(SID, DATA_DIR)
+
+    assert equal(s.data, data)
+
+    # Change data on disk
+    data = randdf()
+    data.to_csv(SUB_PATH / FNAME, index=False)
+
+    # Second time it's cached
+    assert not equal(s.data, data)
+
+    # Clear cache
+    s.clear_cache()
+
+    # Third time it's not cached
+    assert equal(s.data, data)
+
+    # Remove data
+    rmtree(SUB_PATH)
+
+
+# %% Box data container
+
+
+class Box(list):
+    def __init__(self, iterable):
+        super().__init__(iterable)
+
+    def __getattr__(self, name, *args, **kwargs):
+        if hasattr(self[0], name):
+            attr_or_method = getattr(self[0], name)
+            if callable(attr_or_method):
+
+                def fn(*args, **kwargs):
+                    out = []
+                    for elem in self:
+                        result = getattr(elem, name)
+                        result = result(*args, **kwargs) if callable(result) else result
+                        out.append(result)
+                    return out
+
+                return fn
+            else:
+                return [getattr(elem, name) for elem in self]
+        else:
+            raise AttributeError
+
+    def __repr__(self):
+        return f"Box(len={len(self)}, type={self[0].__class__.__module__}.{self[0].__class__.__name__})"
+
+
+# %%
+# Mock class that has attribute access to its underlying data
+df = randdf()
+df.data = df.to_numpy()
+b = Box([df, df])
+b
+
+
+# %%
+b.data
+b.head()
+
+# %%
+box = Box([np.random.randn(10) for i in range(10)])
+
+# %% Run Tests
+test_subject()
+
+# %%
diff --git a/utilz/dftools.py b/utilz/dftools.py
index e18f313..589833d 100644
--- a/utilz/dftools.py
+++ b/utilz/dftools.py
@@ -79,9 +79,10 @@ def _norm(dat, center, scale):
         elif center and scale:
             idx = "normed"
 
+        out = out.to_dict()
         assign_dict = {}
-        for valcol, col in zip(valcols, out):
-            assign_dict[f"{valcol}_{idx}_by_{grpcol}"] = col
+        for key in out.keys():
+            assign_dict[f"{key}_{idx}_by_{grpcol}"] = out[key]
         out = df.assign(**assign_dict)
     return out.squeeze()
 
diff --git a/utilz/genz.py b/utilz/genz.py
new file mode 100644
index 0000000..5bcd389
--- /dev/null
+++ b/utilz/genz.py
@@ -0,0 +1,40 @@
+"""
+    Tools for working with generators. Also see the `toolz` library!
+"""
+
+
+def make_gen(iterable):
+    """Turn any iterable into a generator"""
+    return (e for e in iterable)
+
+
+def combine_gens(*iterables):
+    """
+    Combine multiple generators into a single generator based on every
+    unique combination of their elements. This is equivalent to a series of
+    nested for loops just like itertools.product(). But unlike itertools.product
+    doesn't exhaust each generator before combining them.
+
+    Examples:
+        >>> # This
+        >>> for aa in a:
+        >>>   for bb in b:
+        >>>     for cc in c:
+        >>>       func(aa, bb, cc)
+
+        >>> # Becomes this
+        >>> for aa, bb, cc in combine_gens(a, b, c):
+        >>>   func(aa, bb, cc)
+
+    Yields:
+        generator
+    """
+    if not iterables:
+        # Base case: If no generators are provided, yield an empty tuple.
+        yield ()
+    else:
+        # Recursive case: Combine the first generator with combinations from the rest.
+        first_gen, *remaining_gens = iterables
+        for item in first_gen:
+            for sub_combination in combine_gens(*remaining_gens):
+                yield (item,) + sub_combination
diff --git a/utilz/plot.py b/utilz/plot.py
index 7d19bf8..389df67 100644
--- a/utilz/plot.py
+++ b/utilz/plot.py
@@ -79,6 +79,7 @@ def stripbarplot(
     if pointcolor == "hue":
         ax = sns.stripplot(*args, **kwargs, data=data, ax=ax, alpha=alpha)
     else:
+        _ = kwargs.pop("palette", None)
         ax = sns.stripplot(
             *args, **kwargs, color=pointcolor, data=data, ax=ax, alpha=alpha
         )
@@ -132,6 +133,7 @@ def savefig(
     vector: bool = True,
     use_subdirs: bool = True,
     raster_extension: str = "jpg",
+    vector_extension: str = "svg",
     bbox_inches: str = "tight",
     overwrite: bool = True,
     **kwargs,
@@ -163,10 +165,10 @@ def savefig(
         path = Path.cwd()
     if use_subdirs:
         raster_path = path / "raster" / f"{name}.{raster_extension}"
-        vector_path = path / "vector" / f"{name}.pdf"
+        vector_path = path / "vector" / f"{name}.{vector_extension}"
     else:
         raster_path = path / f"{name}.{raster_extension}"
-        vector_path = path / f"{name}.pdf"
+        vector_path = path / f"{name}.{vector_extension}"
     if not raster_path.parent.exists():
         raster_path.parent.mkdir()
     if not vector_path.parent.exists():
diff --git a/utilz/shorts.py b/utilz/shorts.py
index b7e0eae..a43dc8e 100644
--- a/utilz/shorts.py
+++ b/utilz/shorts.py
@@ -42,23 +42,23 @@ def discard(*args, **kwargs):
     return filter(*args, invert=invert, **kwargs)
 
 
-def seq(n):
+def seq(*args):
     """Enumerated `list`"""
-    return list(range(n))
+    return list(range(*args))
 
 
 def equal(*seqs):
     """
     Checks if N args of potentionally different lengths are equal.
     Non-iterable args are directly compared with `==`
-    Dataframes and arrays use `.equals()` and `np.allclose()` respectively
+    Dataframes and arrays both use `np.allclose()` for comparison
     """
 
     if not isinstance(seqs[0], Iterable):
         return checkall(lambda e: e == seqs[0], seqs)
 
     if isinstance(seqs[0], pd.DataFrame):
-        return checkall(lambda e: e.equals(seqs[0]), seqs)
+        return checkall(lambda e: np.allclose(e.to_numpy(), seqs[0].to_numpy()), seqs)
 
     if isinstance(seqs[0], np.ndarray):
         return checkall(lambda e: np.allclose(e, seqs[0]), seqs)
diff --git a/utilz/tests/test_data.py b/utilz/tests/test_data.py
new file mode 100644
index 0000000..85e9e3c
--- /dev/null
+++ b/utilz/tests/test_data.py
@@ -0,0 +1,102 @@
+import numpy as np
+import pandas as pd
+from utilz.data import Box
+from utilz import randdf, equal, map, seq
+import pytest
+
+
+def test_box():
+    # Dataframe data
+    df_data = [
+        randdf((20, 3), groups={"condition": 2, "group": 4}),
+        randdf((20, 3), groups={"condition": 2, "group": 4}),
+    ]
+
+    box = Box(df_data)
+
+    # We can get the data in a box by slicing it or using .contents()
+    assert box[0].equals(df_data[0])
+    assert all([x.equals(y) for x, y in zip(box[:], df_data)])
+    assert all([x.equals(y) for x, y in zip(box.contents(), df_data)])
+
+    # By default boxes are transparent and always return the contents of an operation
+
+    # Method access
+    out = box.head()
+    assert isinstance(out, list)
+
+    correct = map(lambda x: x.head(), df_data)
+    assert all([x.equals(y) for x, y in zip(out, correct)])
+
+    # Attribute access
+    out = box.shape
+    assert isinstance(out, list)
+
+    correct = map(lambda x: x.shape, df_data)
+    assert all(map(lambda tup: equal(*tup), zip(out, correct)))
+
+    # Numpy arrays
+    data = [np.random.randn(10) for i in range(10)]
+    box = Box(data)
+
+    # Method access
+    out = box.mean()
+    assert isinstance(out, list)
+
+    correct = map(lambda x: x.mean(), data)
+    assert all(map(lambda tup: equal(*tup), zip(out, correct)))
+
+    # Attribute access
+    out = box.shape
+    assert isinstance(out, list)
+
+    correct = map(lambda x: x.shape, data)
+    assert all(map(lambda tup: equal(*tup), zip(out, correct)))
+
+    # Opaque boxes return a new box who's contents can be accessed using .contents()
+    black_box = Box(df_data, transparent=False)
+    assert isinstance(black_box.head(), Box)
+    assert isinstance(black_box.head().contents(), list)
+    # slice notation works too
+    assert isinstance(black_box.head()[:], list)
+
+    # Opaque boxes are useful for method chaining on underlying data
+    out = black_box.groupby("group").mean().contents()
+    assert isinstance(out, list)
+
+    # Doesn't work cause box is transparent
+    with pytest.raises(AttributeError):
+        box = Box(df_data)
+        box.group_by("group").mean()
+
+    # We can change transparency on the fly
+    box.set_transparent(False)
+    compare = box.groupby("group").mean().contents()
+    assert all([x.equals(y) for x, y in zip(out, compare)])
+
+    # Applying arbitrary functions to box elements
+    data = seq(10)
+    correct = seq(1, 11)
+    box = Box(data)
+
+    # By default map returns the result of the operation just like calling other
+    # attributes or methods
+    result = box.map(lambda x: x + 1)
+    assert isinstance(result, list)
+    assert equal(result, correct)
+
+    # Maps respect box transparency which is useful for method chaing
+    box = Box(df_data, transparent=False)
+    result = box.map(lambda x: x["A1"] + 1).head().contents()
+    assert isinstance(result, list)
+
+    # Map operations can also happen inplace which will change box contents without
+    # returning anything.
+    box = Box(data)
+    box.map(lambda x: x + 1, inplace=True)
+    assert equal(box.contents(), correct)
+
+    # In place doesn't care if the box is transparent or not.
+    box = Box(data, transparent=False)
+    box.map(lambda x: x + 1, inplace=True)
+    assert equal(box.contents(), correct)
diff --git a/utilz/tests/test_genz.py b/utilz/tests/test_genz.py
new file mode 100644
index 0000000..61978bb
--- /dev/null
+++ b/utilz/tests/test_genz.py
@@ -0,0 +1,20 @@
+from types import GeneratorType
+from utilz.genz import make_gen, combine_gens
+import pytest
+
+
+def test_make_gen():
+    l = list(range(10))
+    g = make_gen(l)
+    assert isinstance(g, GeneratorType)
+    breakpoint()
+
+
+def test_combine_gens():
+    l1 = list(range(10))
+    l2 = list(range(5))
+    l3 = list(range(3))
+    l4 = list(range(20))
+    out = list(combine_gens(l1, l2, l3, l4))
+    assert len(out) == len(l1) * len(l2) * len(l3) * len(l4)
+    assert out[-1] == (9, 4, 2, 19)
diff --git a/utilz/tests/test_ops.py b/utilz/tests/test_ops.py
index 7252617..0490216 100644
--- a/utilz/tests/test_ops.py
+++ b/utilz/tests/test_ops.py
@@ -23,7 +23,6 @@
     discard,
     seq,
     equal,
-    fork,
 )
 from utilz.plot import tweak
 from utilz.boilerplate import randdf
diff --git a/utilz/tests/test_plot.py b/utilz/tests/test_plot.py
index f2391c7..01af78c 100644
--- a/utilz/tests/test_plot.py
+++ b/utilz/tests/test_plot.py
@@ -28,7 +28,7 @@ def test_savefig(tmp_path: Path):
     # Save to cwd
     f, _ = mpinit(subplots=(2, 2))
     save_raster = Path.cwd() / "raster" / "test.jpg"
-    save_vector = Path.cwd() / "vector" / "test.pdf"
+    save_vector = Path.cwd() / "vector" / "test.svg"
     savefig(f, "test")
     plt.close(f)
     assert save_raster.exists()
@@ -41,7 +41,7 @@ def test_savefig(tmp_path: Path):
     # Save to custom path
     tmp_path = Path(tmp_path)
     dir_save_raster = tmp_path / "test.jpg"
-    dir_save_vector = tmp_path / "test.pdf"
+    dir_save_vector = tmp_path / "test.svg"
 
     f, _ = mpinit(subplots=(2, 2))
     savefig(f, "test", path=tmp_path, use_subdirs=False)