Desire and Delaunay lines

Creating Desire and Delaunay Lines

Creating county-to-county matrices

sphinx_gallery_thumbnail_path = ‘../../examples/result_analysis/delaunay_lines.png’

from pathlib import Path

import numpy as np

from polaris.analyze.mapping.flow_lines import delaunay_assignment, desire_lines
from polaris.analyze.trip_metrics import TripMetrics

Polaris Studio enables you to create Desire and Delaunay lines for trip matrices between counties or between zones. Let’s show that using the Bloomington model as an example.

project_dir = Path("/tmp/Bloomington")

last_iter = TripMetrics(project_dir / "Bloomington-Supply.sqlite", project_dir / "Bloomington-Demand.sqlite")

# Let's get all the trips for the last iteration
matrix_zones = last_iter.trip_matrix(from_start_time=0, to_start_time=24 * 3600, aggregation="zone")

# And let's see what modes we have:
matrix_zones.matrices

['BUS', 'SOV', 'TAXI']

We can plot the zone-to-zone flows as Delaunay lines

flows is a GeoDataFrame, so you can save it to a file for use elsewhere if you’d like

flows = delaunay_assignment(project_dir / "Bloomington-Supply.sqlite", aggregation="zone", matrix=matrix_zones)

flows.explore(
    style_kwds={
        "style_function": lambda x: {
            "weight": x["properties"]["SOV_tot"] / flows["SOV_tot"].max() * 10  # scale as needed
        }
    },
    tiles="CartoDB positron",
    attr="Map tiles by Stamen Design, CC BY 3.0 — Map data © OpenStreetMap contributors",
)

delaunay                                          :   0%|          | 0/185 [00:00<?, ?it/s]
Equilibrium Assignment                            :   0%|          | 0/250 [00:00<?, ?it/s]

Make this Notebook Trusted to load map: File -> Trust Notebook

Desire lines

To mix it up a little, we can plot desire lines.

dls = desire_lines(project_dir / "Bloomington-Supply.sqlite", aggregation="zone", matrix=matrix_zones)

But we plot only flows larger than 50 trips, or the map will be impossibly heavy to render/navigate

dls2 = dls[dls["SOV_tot"] >= np.partition(dls["SOV_tot"].to_numpy(), -50)[-50]]
dls2.explore(
    style_kwds={
        "style_function": lambda x: {
            "weight": x["properties"]["SOV_tot"] / dls["SOV_tot"].max() * 20  # scale as needed
        }
    },
    tiles="CartoDB positron",
    attr="Map tiles by Stamen Design, CC BY 3.0 — Map data © OpenStreetMap contributors",
)

Make this Notebook Trusted to load map: File -> Trust Notebook

You can always simply plot the GeoDataFrame using the plot method, but it will not be interactive.

dls.plot(linewidth=5 * dls["SOV_tot"] / dls["SOV_tot"].max(), color="red")

<Axes: >

# If we wanted, we could've done this all for county flows too

# matrix_counties = last_iter.trip_matrix(from_start_time=0, to_start_time=24 * 3600, aggregation="county")

# dls = desire_lines(project_dir / "Bloomington-Supply.sqlite", aggregation="county", matrix=matrix_counties)

# flows = delaunay_assignment(project_dir / "Bloomington-Supply.sqlite", aggregation="county", matrix=matrix_counties)