1 of 10

CSE 163

Dissolve and Join�

Suh Young Choi�

2 of 10

Announcements

Checkpoint 4 due on Monday

THA 5: Mapping due on Thursday

New Resubmission cycle!

THA 4 can be resubmitted

2

3 of 10

This Time

Dissolve (groupby for Geopandas)
Joining datasets

Last Time

Geospatial Data
Using Geopandas

3

4 of 10

Geopandas

Exactly like pandas, but with some extra features for geo data
Types

GeoDataFrame
GeoSeries

4

import geopandas as gpd

df = gpd.read_file('data_file.shp')

df.plot(column='some_col', legend=True)

plt.savefig('plot.png')

5 of 10

Matplotlib

Before, we were using a lot of matplotlib stuff blindly
Default plots gets plotted on a global figure (think canvas)

Multiple plots would overwrite each other

You have to do extra work to get more information on the same set of axes.

5

fig, ax = plt.subplots(1)

<plot1>(ax=ax)

<plot2>(ax=ax)

fig.savefig('plot.png')

A few reminders about Matplotlib. Before, we were using a lot of mpl stuff without really thinking about what it meant. By default, any plots you have are plotted on a global figure, which is like a giant blank canvas where your plots live. By default, multiple plots override one another, which is why it’s important to save figures between calls to plotting functions in Seaborn. We need a bit of extra work in order to get information on the same set of axes. This is like the GDP rpbolem that we saw on Monday, where we wanted to plot the entire map of the world, and then some subset of the countries on top of that. To make that work, we need to specify both a figure and axis (or multiple axes, if you want plots side-y-side.) To get multiple layers on the same plot, we end up using this ax=ax syntax. Reminder that the first “ax” is the name of the parameter, and the second “Ax” is the literal ax figure that we defined above.

6 of 10

Dissolve

Exactly the same as a groupby for the the regular columns

For the geometry columns, overlays all of the geometries

Options for aggfunc

‘first’
‘last’
‘min’
‘max’

6

7 of 10

Join

7

tas.merge(grading, left_on='ta_id',

right_on='grader_id')

ta_name	ta_id
Ryan	1
James	2
Nicole	3

grader_id	student_name
2	Flora
3	Paul
1	Wen
3	Andrew

ta_name	ta_id	grader_id	student_name
Ryan	1	1	Wen
James	2	2	Flora
Nicole	3	3	Paul
Nicole	3	3	Andrew

8 of 10

Types of Join

There are interesting questions of what happens if there are rows that don’t “line up”. Different type of joins differ in how to handle this case.

Types of Joins

left.merge(right, left_on=’lcol’, right_on=’rcol’,

how=’type’)

Inner (default): Both values must be present
Left: If a value from left has no match, add NaNs
Right: If a value from right has no match, add NaNs
Outer: If a value from either table has no match, add NaNs

8

9 of 10

Hurricane Florence

9

10 of 10

Before Next Time

Complete Lesson 18

Remember not for points, but do go towards Weekly Tokens

Complete Checkpoint 4!
Keep working on THA 5!
See you next time :)

Next Time

numpy
Images

10