Spatial join is yet another classic GIS problem. Getting attributes from one layer and transferring them into another layer based on their spatial relationship is something you most likely need to do on a regular basis.
Sources
Following materials are partly based on documentation of Geopandas.
The previous materials focused on learning how to perform a Point in Polygon query. We could now apply those techniques and create our own function to perform a spatial join between two layers based on their spatial relationship. We could for example join the attributes of a polygon layer into a point layer where each point would get the attributes of a polygon that contains the point.
Luckily, spatial join (gpd.sjoin() -function) is already implemented in Geopandas, thus we do not need to create it ourselves. There are three possible types of join that can be applied in spatial join that are determined with op -parameter:
"intersects""within""contains"Sounds familiar? Yep, all of those spatial relationships were discussed in the previous materials, thus you should know how they work.
Let's perform a spatial join between the address-point Shapefile that we created and then reprojected and a Polygon layer that is a 250m x 250m grid showing the amount of people living in Helsinki Region.
For this lesson we will be using publicly available population data from Helsinki that can be downloaded from Helsinki Region Infroshare (HRI) which is an excellent source that provides all sorts of open data from Helsinki, Finland. From HRI download a Population grid for year 2015/V%C3%A4est%C3%B6tietoruudukko/Vaestotietoruudukko_2015.zip) that is a dataset (.shp) produced by Helsinki Region Environmental Services Authority (HSY) (see this page to access data from different years).
$ cd
$ unzip Vaestotietoruudukko_2015.zip -d Pop15
$ ls Pop15
Vaestotietoruudukko_2015.dbf Vaestotietoruudukko_2015.shp
Vaestotietoruudukko_2015.prj Vaestotietoruudukko_2015.shx
You should now have a folder /home/geo/Pop15 with files listed above.
In [3]:
import geopandas as gpd
# Filepath
fp = r"/home/geo/Pop15/Vaestotietoruudukko_2015.shp"
# Read the data
pop = gpd.read_file(fp)
# See the first rows
print(pop.head())
Okey so we have multiple columns in the dataset but the most important one here is the column ASUKKAITA (population in Finnish) that tells the amount of inhabitants living under that polygon.
pop15 so that it is more intuitive. Changing column names is easy in Pandas / Geopandas using a function called rename() where we pass a dictionary to a parameter columns={'oldname': 'newname'}.
In [4]:
# Change the name of a column
pop = pop.rename(columns={'ASUKKAITA': 'pop15'})
# See the column names and confirm that we now have a column called 'pop15'
print(pop.columns)
pop15 and geometry
In [5]:
# Columns that will be sected
selected_cols = ['pop15', 'geometry']
# Select those columns
pop = pop[selected_cols]
# Let's see the last 2 rows
print(pop.tail(2))
Now we have cleaned the data and have only those columns that we need for our analysis.
Now we are ready to perform the spatial join between the two layers that we have. The aim here is to get information about how many people live in a polygon that contains an individual address-point . Thus, we want to join attributes from the population layer we just modified into the addresses point layer addresses_epsg3879.shp.
In [12]:
# Addresses filpath
addr_fp = r"/home/geo/addresses_epsg3879.shp"
# Read data
addresses = gpd.read_file(addr_fp)
# Check the head of the file
print(addresses.head(2))
In [13]:
# Check the crs of address points
print(addresses.crs)
# Check the crs of population layer
print(pop.crs)
# Do they match? - We can test that
addresses.crs == pop.crs
Out[13]:
Indeed they are identical. Thus, we can be sure that when doing spatial queries between layers the locations match and we get the right results e.g. from the spatial join that we are conducting here.
pop GeoDataFrame into addresses GeoDataFrame by using gpd.sjoin() -function
In [43]:
# Make a spatial join
join = gpd.sjoin(addresses, pop, how="inner", op="within")
# Let's check the result
print(join.head())
Awesome! Now we have performed a successful spatial join where we got two new columns into our join GeoDataFrame, i.e. index_right that tells the index of the matching polygon in the pop layer and pop15 which is the population in the cell where the address-point is located.
In [44]:
# Output path
outfp = r"/home/geo/addresses_pop15_epsg3979.shp"
# Save to disk
join.to_file(outfp)
Do the results make sense? Let's evaluate this a bit by plotting the points where color intensity indicates the population numbers.
pop15 column to indicate the color. cmap -parameter tells to use a sequential colormap for the values, markersize adjusts the size of a point, scheme parameter can be used to adjust the classification method based on pysal, and legend tells that we want to have a legend.
In [3]:
import matplotlib.pyplot as plt
join.plot(column='pop15', cmap="Reds", markersize=7, scheme='natural_breaks', legend=True)
By knowing approximately how population is distributed in Helsinki, it seems that the results do make sense as the points with highest population are located in the south where the city center of Helsinki is.