Blog - overview

_posts/2018-08-15-plotting-overview.md

@@ -0,0 +1,84 @@
+---
+layout: post
+title:  "Altair vs Matplotlib Plot Construction Overview"
+date:   2018-08-15 14:50:00 -0500
+author: "Kimberly Orr"
+categories: about
+tags: "intro about matplotlib altair"
+excerpt_separator: <!--read more-->
+---
+
+# Plot Construction Overview
+We have data that we want to plot!
+```python
+df = pd.DataFrame({
+    'x': [1, 2, 3, 4], 'y': [8, 7, 6, 5], 'c': [1, 1, 2, 3]
+})
+```
+Fortunately, python gives us options for plotting that data. In this post, we'll look at plotting this data with Altair, Matplotlib, and mpl-altair.
+
+## Altair
+Altair is a Python visualization library built on top of the Vega/Vegalite declarative grammar. As Altair's [overview](https://altair-viz.github.io/getting_started/overview.html) states:
+>The key idea is that you are declaring links between data columns and visual encoding channels, such as the x-axis, y-axis, color, etc.
+
+So, the general process for creating a basic Altair plot is to specify your data:
+ ```python
+alt.Chart(df)
+ ``` 
+Specify what type of glyph/marker should be used to represent your data:
+```python
+.mark_point()
+```
+Then link your data columns with the encoding channels:
+ ```python
+.encode(x=alt.X('x'), y=alt.Y('y'), color=alt.Color('c'))
+ ```
+
+So that a finished plot would look like:
+ ```python
+# import
+import altair as alt
+# plot
+alt.Chart(df).mark_point().encode(
+    x=alt.X('x'), y=alt.Y('y'), color=alt.Color('c')
+)
+```
+## Matplotlib
+Matplotlib is a powerful object-oriented procedural plotting library. The general process to create a Matplotlib plot is to first create a figure (canvas) and one or more subplots (which are objects that encapsulate plot data to facilitate adding multiple plots to a single canvas) and then add objects to the subplots (like axes, which contain glyphs, etc.). So, instead of linking data with encoding channels, Matplotlib uses an object-oriented interface to place objects on a canvas.
+
+The general thought process is to create the canvas/subplots:
+```python
+fig, ax = plt.subplots()
+```
+Add a scatter plot to the axes object of this figure:
+```python
+ax.scatter(df['x'].values, df['y'].values, c=df['c'].values)
+```
+Show it:
+```python
+plt.show()
+```
+
+So that a plot of `y_array` vs `x_array` colored by `color_array` would look like this:
+```python
+import matplotlib.pyplot as plt
+# plot
+fix, ax = plt.subplots()
+ax.scatter(df['x'].values, df['y'].values, c=df['c'].values)
+plt.show()
+```
+
+## mpl-altair
+mpl-altair allows you to create an altair chart as normal and convert/render it as a Matplotlib figure like so:
+```python
+import altair as alt
+import matplotlib.pyplot as plt
+import mplaltair
+# make an altair chart
+chart = alt.Chart(df).mark_point().encode(
+    alt.X('x'), alt.Y('y'), alt.Color('c')
+)
+# convert to Matplotlib
+fig, ax = mplaltair.convert(chart)
+plt.show()
+```