Modifying the population using complex policies¶

This notebook shows how complex policies can be applied to people, households, and entire populations.

In [1]:

from pam.activity import Activity, Leg
from pam.core import Household, Person, Population
from pam.policy import (
    ActivityPolicy,
    ActivityProbability,
    HouseholdPolicy,
    MoveActivityTourToHomeLocation,
    PersonAttributeFilter,
    PersonPolicy,
    PersonProbability,
    ReduceSharedActivity,
    RemoveActivity,
    apply_policies,
)
from pam.utils import minutes_to_datetime as mtdt
from pam.variables import END_OF_DAY

from pam.activity import Activity, Leg from pam.core import Household, Person, Population from pam.policy import ( ActivityPolicy, ActivityProbability, HouseholdPolicy, MoveActivityTourToHomeLocation, PersonAttributeFilter, PersonPolicy, PersonProbability, ReduceSharedActivity, RemoveActivity, apply_policies, ) from pam.utils import minutes_to_datetime as mtdt from pam.variables import END_OF_DAY

In [2]:

def print_simple_stats(population):
    """Print some simple population statistics."""
    time_at_home = 0
    travel_time = 0
    low_income_central_trips = 0
    high_income_central_trips = 0

    for _hid, hh in population.households.items():
        for _pid, person in hh.people.items():
            freq = person.freq

            for p in person.plan:
                if p.act == "travel":
                    duration = p.duration.seconds * freq / 3600
                    travel_time += duration

                    if p.end_location.area == "Westminster,City of London":
                        if person.attributes["inc"] == "low":
                            low_income_central_trips += freq

                        elif person.attributes["inc"] == "high":
                            high_income_central_trips += freq

                else:  # activity
                    if p.act == "home":
                        duration = p.duration.seconds * freq / 3600
                        time_at_home += duration

    print(f"Population total time at home: {time_at_home:.2f} hours")
    print(f"Population total travel time: {travel_time:.2f} hours")

def print_simple_stats(population): """Print some simple population statistics.""" time_at_home = 0 travel_time = 0 low_income_central_trips = 0 high_income_central_trips = 0 for _hid, hh in population.households.items(): for _pid, person in hh.people.items(): freq = person.freq for p in person.plan: if p.act == "travel": duration = p.duration.seconds * freq / 3600 travel_time += duration if p.end_location.area == "Westminster,City of London": if person.attributes["inc"] == "low": low_income_central_trips += freq elif person.attributes["inc"] == "high": high_income_central_trips += freq else: # activity if p.act == "home": duration = p.duration.seconds * freq / 3600 time_at_home += duration print(f"Population total time at home: {time_at_home:.2f} hours") print(f"Population total travel time: {travel_time:.2f} hours")

Create example Population¶

In [3]:

population = Population()

population = Population()

Add Smith household¶

In [4]:

smith_id = "Smith"

Billy = Person(
    "Billy",
    freq=1,
    attributes={"age": 26, "job": "employed", "gender": "female", "key_worker": False},
)

Billy.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60)))
Billy.add(
    Leg(1, "car", start_area="a", end_area="g", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 20))
)
Billy.add(Activity(2, "work", "g", start_time=mtdt(8 * 60 + 20), end_time=mtdt(17 * 60)))
Billy.add(
    Leg(
        2,
        "car",
        start_area="g",
        end_area="a",
        start_time=mtdt(17 * 60),
        end_time=mtdt(17 * 60 + 25),
    )
)
Billy.add(Activity(3, "home", "a", start_time=mtdt(17 * 60 + 25), end_time=END_OF_DAY))


Bobby = Person(
    "Bobby",
    freq=1,
    attributes={"age": 6, "job": "education", "gender": "male", "key_worker": False},
)
Bobby.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60)))
Bobby.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60 + 5), end_time=mtdt(8 * 60 + 30)))
Bobby.add(Activity(2, "education", "b", start_time=mtdt(8 * 60 + 30), end_time=mtdt(16 * 60)))
Bobby.add(Leg(2, "walk", "b", "c", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 35)))
Bobby.add(Activity(3, "home", "a", start_time=mtdt(16 * 60 + 30), end_time=mtdt(18 * 60)))
Bobby.add(Leg(3, "car", "a", "b", start_time=mtdt(18 * 60), end_time=mtdt(18 * 60 + 20)))
Bobby.add(Activity(4, "shop_1", "b", start_time=mtdt(18 * 60 + 20), end_time=mtdt(18 * 60 + 50)))
Bobby.add(Leg(4, "car", "b", "b", start_time=mtdt(18 * 60 + 50), end_time=mtdt(19 * 60)))
Bobby.add(Activity(5, "shop_2", "b", start_time=mtdt(19 * 60), end_time=mtdt(19 * 60 + 50)))
Bobby.add(Leg(5, "car", "b", "a", start_time=mtdt(19 * 60 + 50), end_time=mtdt(20 * 60 + 10)))
Bobby.add(Activity(6, "home", "a", start_time=mtdt(20 * 60 + 10), end_time=END_OF_DAY))

Bradly = Person(
    "Bradly",
    freq=1,
    attributes={"age": 40, "job": "employed", "gender": "male", "key_worker": True},
)
Bradly.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60)))

Bradly.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 20)))
Bradly.add(Activity(2, "escort", "b", start_time=mtdt(8 * 60 + 20), end_time=mtdt(8 * 60 + 30)))
Bradly.add(Leg(2, "pt", "b", "b", start_time=mtdt(8 * 60 + 30), end_time=mtdt(9 * 60)))

Bradly.add(Activity(4, "work", "b", start_time=mtdt(9 * 60), end_time=mtdt(10 * 60)))
Bradly.add(Leg(3, "pt", "b", "c", start_time=mtdt(10 * 60), end_time=mtdt(10 * 60 + 20)))
Bradly.add(Activity(4, "work", "b", start_time=mtdt(10 * 60 + 20), end_time=mtdt(12 * 60)))
Bradly.add(Leg(4, "pt", "b", "c", start_time=mtdt(12 * 60), end_time=mtdt(12 * 60 + 20)))
Bradly.add(Activity(5, "work", "b", start_time=mtdt(12 * 60 + 20), end_time=mtdt(14 * 60)))
Bradly.add(Leg(5, "pt", "b", "c", start_time=mtdt(14 * 60), end_time=mtdt(14 * 60 + 20)))

Bradly.add(Activity(6, "leisure", "c", start_time=mtdt(14 * 60 + 20), end_time=mtdt(15 * 60 + 30)))
Bradly.add(Leg(4, "pt", "c", "b", start_time=mtdt(15 * 60 + 30), end_time=mtdt(16 * 60 - 10)))
Bradly.add(Activity(5, "escort", "b", start_time=mtdt(16 * 60 - 10), end_time=mtdt(16 * 60)))
Bradly.add(Leg(5, "walk", "b", "a", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 20)))

Bradly.add(Activity(8, "home", "a", start_time=mtdt(16 * 60 + 20), end_time=mtdt(18 * 60)))
Bradly.add(Leg(8, "car", "a", "b", start_time=mtdt(18 * 60), end_time=mtdt(18 * 60 + 20)))
Bradly.add(Activity(9, "shop_1", "b", start_time=mtdt(18 * 60 + 20), end_time=mtdt(18 * 60 + 50)))
Bradly.add(Leg(9, "car", "b", "b", start_time=mtdt(18 * 60 + 50), end_time=mtdt(19 * 60)))
Bradly.add(Activity(10, "shop_2", "b", start_time=mtdt(19 * 60), end_time=mtdt(19 * 60 + 50)))
Bradly.add(Leg(10, "car", "b", "a", start_time=mtdt(19 * 60 + 50), end_time=mtdt(20 * 60 + 10)))
Bradly.add(Activity(11, "home", "a", start_time=mtdt(20 * 60 + 10), end_time=END_OF_DAY))

smiths = Household(smith_id)
for person in [Billy, Bradly, Bobby]:
    smiths.add(person)

smiths.people

smith_id = "Smith" Billy = Person( "Billy", freq=1, attributes={"age": 26, "job": "employed", "gender": "female", "key_worker": False}, ) Billy.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60))) Billy.add( Leg(1, "car", start_area="a", end_area="g", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 20)) ) Billy.add(Activity(2, "work", "g", start_time=mtdt(8 * 60 + 20), end_time=mtdt(17 * 60))) Billy.add( Leg( 2, "car", start_area="g", end_area="a", start_time=mtdt(17 * 60), end_time=mtdt(17 * 60 + 25), ) ) Billy.add(Activity(3, "home", "a", start_time=mtdt(17 * 60 + 25), end_time=END_OF_DAY)) Bobby = Person( "Bobby", freq=1, attributes={"age": 6, "job": "education", "gender": "male", "key_worker": False}, ) Bobby.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60))) Bobby.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60 + 5), end_time=mtdt(8 * 60 + 30))) Bobby.add(Activity(2, "education", "b", start_time=mtdt(8 * 60 + 30), end_time=mtdt(16 * 60))) Bobby.add(Leg(2, "walk", "b", "c", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 35))) Bobby.add(Activity(3, "home", "a", start_time=mtdt(16 * 60 + 30), end_time=mtdt(18 * 60))) Bobby.add(Leg(3, "car", "a", "b", start_time=mtdt(18 * 60), end_time=mtdt(18 * 60 + 20))) Bobby.add(Activity(4, "shop_1", "b", start_time=mtdt(18 * 60 + 20), end_time=mtdt(18 * 60 + 50))) Bobby.add(Leg(4, "car", "b", "b", start_time=mtdt(18 * 60 + 50), end_time=mtdt(19 * 60))) Bobby.add(Activity(5, "shop_2", "b", start_time=mtdt(19 * 60), end_time=mtdt(19 * 60 + 50))) Bobby.add(Leg(5, "car", "b", "a", start_time=mtdt(19 * 60 + 50), end_time=mtdt(20 * 60 + 10))) Bobby.add(Activity(6, "home", "a", start_time=mtdt(20 * 60 + 10), end_time=END_OF_DAY)) Bradly = Person( "Bradly", freq=1, attributes={"age": 40, "job": "employed", "gender": "male", "key_worker": True}, ) Bradly.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60))) Bradly.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 20))) Bradly.add(Activity(2, "escort", "b", start_time=mtdt(8 * 60 + 20), end_time=mtdt(8 * 60 + 30))) Bradly.add(Leg(2, "pt", "b", "b", start_time=mtdt(8 * 60 + 30), end_time=mtdt(9 * 60))) Bradly.add(Activity(4, "work", "b", start_time=mtdt(9 * 60), end_time=mtdt(10 * 60))) Bradly.add(Leg(3, "pt", "b", "c", start_time=mtdt(10 * 60), end_time=mtdt(10 * 60 + 20))) Bradly.add(Activity(4, "work", "b", start_time=mtdt(10 * 60 + 20), end_time=mtdt(12 * 60))) Bradly.add(Leg(4, "pt", "b", "c", start_time=mtdt(12 * 60), end_time=mtdt(12 * 60 + 20))) Bradly.add(Activity(5, "work", "b", start_time=mtdt(12 * 60 + 20), end_time=mtdt(14 * 60))) Bradly.add(Leg(5, "pt", "b", "c", start_time=mtdt(14 * 60), end_time=mtdt(14 * 60 + 20))) Bradly.add(Activity(6, "leisure", "c", start_time=mtdt(14 * 60 + 20), end_time=mtdt(15 * 60 + 30))) Bradly.add(Leg(4, "pt", "c", "b", start_time=mtdt(15 * 60 + 30), end_time=mtdt(16 * 60 - 10))) Bradly.add(Activity(5, "escort", "b", start_time=mtdt(16 * 60 - 10), end_time=mtdt(16 * 60))) Bradly.add(Leg(5, "walk", "b", "a", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 20))) Bradly.add(Activity(8, "home", "a", start_time=mtdt(16 * 60 + 20), end_time=mtdt(18 * 60))) Bradly.add(Leg(8, "car", "a", "b", start_time=mtdt(18 * 60), end_time=mtdt(18 * 60 + 20))) Bradly.add(Activity(9, "shop_1", "b", start_time=mtdt(18 * 60 + 20), end_time=mtdt(18 * 60 + 50))) Bradly.add(Leg(9, "car", "b", "b", start_time=mtdt(18 * 60 + 50), end_time=mtdt(19 * 60))) Bradly.add(Activity(10, "shop_2", "b", start_time=mtdt(19 * 60), end_time=mtdt(19 * 60 + 50))) Bradly.add(Leg(10, "car", "b", "a", start_time=mtdt(19 * 60 + 50), end_time=mtdt(20 * 60 + 10))) Bradly.add(Activity(11, "home", "a", start_time=mtdt(20 * 60 + 10), end_time=END_OF_DAY)) smiths = Household(smith_id) for person in [Billy, Bradly, Bobby]: smiths.add(person) smiths.people

Out[4]:

{'Billy': <pam.core.Person at 0x106c11600>,
 'Bradly': <pam.core.Person at 0x137e48ac0>,
 'Bobby': <pam.core.Person at 0x137e35c60>}

In [5]:

smiths.plot()

smiths.plot()

In [6]:

population.add(smiths)

population.add(smiths)

Add Jones household¶

In [7]:

jones_id = "Jones"

Hugh = Person(
    "Hugh",
    freq=1,
    attributes={"age": 100, "job": "unemployed", "gender": "male", "key_worker": False},
)
Hugh.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60)))
Hugh.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 30)))
Hugh.add(Activity(2, "health", "b", start_time=mtdt(8 * 60 + 30), end_time=mtdt(10 * 60)))
Hugh.add(Leg(2, "walk", "b", "a", start_time=mtdt(10 * 60), end_time=mtdt(10 * 60 + 30)))
Hugh.add(Activity(3, "home", "a", start_time=mtdt(10 * 60 + 30), end_time=mtdt(14 * 60)))
Hugh.add(Leg(3, "walk", "a", "b", start_time=mtdt(14 * 60), end_time=mtdt(14 * 60 + 30)))
Hugh.add(Activity(4, "health", "b", start_time=mtdt(14 * 60 + 30), end_time=mtdt(16 * 60)))
Hugh.add(Leg(4, "walk", "b", "a", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 30)))
Hugh.add(Activity(5, "home", "a", start_time=mtdt(16 * 60 + 30), end_time=END_OF_DAY))

Bridget = Person(
    "Bridget",
    freq=1,
    attributes={"age": 35, "job": "employed", "gender": "female", "key_worker": False},
)
Bridget.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60)))
Bridget.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 5)))
Bridget.add(Activity(2, "escort", "b", start_time=mtdt(8 * 60 + 5), end_time=mtdt(10 * 60 + 30)))
Bridget.add(Leg(2, "pt", "b", "c", start_time=mtdt(10 * 60 + 30), end_time=mtdt(11 * 60)))
Bridget.add(Activity(3, "work", "c", start_time=mtdt(11 * 60), end_time=mtdt(16 * 60)))
Bridget.add(Leg(3, "pt", "c", "a", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 20)))
Bridget.add(Activity(4, "home", "a", start_time=mtdt(16 * 60 + 20), end_time=mtdt(17 * 60 + 20)))
Bridget.add(Leg(4, "pt", "c", "a", start_time=mtdt(17 * 60 + 20), end_time=mtdt(17 * 60 + 50)))
Bridget.add(Activity(5, "shop", "a", start_time=mtdt(17 * 60 + 50), end_time=mtdt(18 * 60 + 30)))
Bridget.add(Leg(5, "pt", "c", "a", start_time=mtdt(18 * 60 + 30), end_time=mtdt(18 * 60 + 50)))
Bridget.add(Activity(6, "home", "a", start_time=mtdt(18 * 60 + 50), end_time=END_OF_DAY))

jones = Household(jones_id)
for person in [Hugh, Bridget]:
    jones.add(person)

jones.people

jones_id = "Jones" Hugh = Person( "Hugh", freq=1, attributes={"age": 100, "job": "unemployed", "gender": "male", "key_worker": False}, ) Hugh.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60))) Hugh.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 30))) Hugh.add(Activity(2, "health", "b", start_time=mtdt(8 * 60 + 30), end_time=mtdt(10 * 60))) Hugh.add(Leg(2, "walk", "b", "a", start_time=mtdt(10 * 60), end_time=mtdt(10 * 60 + 30))) Hugh.add(Activity(3, "home", "a", start_time=mtdt(10 * 60 + 30), end_time=mtdt(14 * 60))) Hugh.add(Leg(3, "walk", "a", "b", start_time=mtdt(14 * 60), end_time=mtdt(14 * 60 + 30))) Hugh.add(Activity(4, "health", "b", start_time=mtdt(14 * 60 + 30), end_time=mtdt(16 * 60))) Hugh.add(Leg(4, "walk", "b", "a", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 30))) Hugh.add(Activity(5, "home", "a", start_time=mtdt(16 * 60 + 30), end_time=END_OF_DAY)) Bridget = Person( "Bridget", freq=1, attributes={"age": 35, "job": "employed", "gender": "female", "key_worker": False}, ) Bridget.add(Activity(1, "home", "a", start_time=mtdt(0), end_time=mtdt(8 * 60))) Bridget.add(Leg(1, "walk", "a", "b", start_time=mtdt(8 * 60), end_time=mtdt(8 * 60 + 5))) Bridget.add(Activity(2, "escort", "b", start_time=mtdt(8 * 60 + 5), end_time=mtdt(10 * 60 + 30))) Bridget.add(Leg(2, "pt", "b", "c", start_time=mtdt(10 * 60 + 30), end_time=mtdt(11 * 60))) Bridget.add(Activity(3, "work", "c", start_time=mtdt(11 * 60), end_time=mtdt(16 * 60))) Bridget.add(Leg(3, "pt", "c", "a", start_time=mtdt(16 * 60), end_time=mtdt(16 * 60 + 20))) Bridget.add(Activity(4, "home", "a", start_time=mtdt(16 * 60 + 20), end_time=mtdt(17 * 60 + 20))) Bridget.add(Leg(4, "pt", "c", "a", start_time=mtdt(17 * 60 + 20), end_time=mtdt(17 * 60 + 50))) Bridget.add(Activity(5, "shop", "a", start_time=mtdt(17 * 60 + 50), end_time=mtdt(18 * 60 + 30))) Bridget.add(Leg(5, "pt", "c", "a", start_time=mtdt(18 * 60 + 30), end_time=mtdt(18 * 60 + 50))) Bridget.add(Activity(6, "home", "a", start_time=mtdt(18 * 60 + 50), end_time=END_OF_DAY)) jones = Household(jones_id) for person in [Hugh, Bridget]: jones.add(person) jones.people

Out[7]:

{'Hugh': <pam.core.Person at 0x16424cd00>,
 'Bridget': <pam.core.Person at 0x1641253c0>}

In [8]:

jones.plot()

jones.plot()

In [9]:

population.add(jones)

population.add(jones)

In [10]:

print_simple_stats(population)

print_simple_stats(population)

Population total time at home: 74.58 hours
Population total travel time: 9.67 hours

PAM Complex Policies¶

Based on link

In [11]:

def discrete_joint_distribution_sampler(obj, mapping, distribution):
    """
    Randomly sample from a joint distribution based some discrete features.
    Where features are a dictionary structure of features, eg: {'gender':'female'}
    Distribution is a nested dict of probabilities based on possible features, eg:
    {'0-0': {'male': 0, 'female': 0},... , '90-120': {'male': 1, 'female': 1}}
    Mapping provides the feature name for each level of the distribution, eg:
    ['age', 'gender'].
    """
    p = distribution
    for key in mapping:
        value = obj.attributes.get(key)
        if value is None:
            msg = f"Cannot find mapping: {key} in sampling features: {obj.attributes}"
            raise KeyError(msg)
        p = p.get(value)
        if p is None:
            msg = f"Cannot find feature for {key}: {value} in distribution: {p}"
            raise KeyError(msg)

    return p

def discrete_joint_distribution_sampler(obj, mapping, distribution): """ Randomly sample from a joint distribution based some discrete features. Where features are a dictionary structure of features, eg: {'gender':'female'} Distribution is a nested dict of probabilities based on possible features, eg: {'0-0': {'male': 0, 'female': 0},... , '90-120': {'male': 1, 'female': 1}} Mapping provides the feature name for each level of the distribution, eg: ['age', 'gender']. """ p = distribution for key in mapping: value = obj.attributes.get(key) if value is None: msg = f"Cannot find mapping: {key} in sampling features: {obj.attributes}" raise KeyError(msg) p = p.get(value) if p is None: msg = f"Cannot find feature for {key}: {value} in distribution: {p}" raise KeyError(msg) return p

In [12]:

vulnerable_mapping = ["age", "gender"]
vulnerable_distribution = dict(
    zip(
        list(range(101)),
        [{"male": i / 100, "female": i / 100, "other": i / 100} for i in range(101)],
    )
)

vulnerable_mapping = ["age", "gender"] vulnerable_distribution = dict( zip( list(range(101)), [{"male": i / 100, "female": i / 100, "other": i / 100} for i in range(101)], ) )

In [13]:

dict(list(vulnerable_distribution.items())[0:15])

dict(list(vulnerable_distribution.items())[0:15])

Out[13]:

{0: {'male': 0.0, 'female': 0.0, 'other': 0.0},
 1: {'male': 0.01, 'female': 0.01, 'other': 0.01},
 2: {'male': 0.02, 'female': 0.02, 'other': 0.02},
 3: {'male': 0.03, 'female': 0.03, 'other': 0.03},
 4: {'male': 0.04, 'female': 0.04, 'other': 0.04},
 5: {'male': 0.05, 'female': 0.05, 'other': 0.05},
 6: {'male': 0.06, 'female': 0.06, 'other': 0.06},
 7: {'male': 0.07, 'female': 0.07, 'other': 0.07},
 8: {'male': 0.08, 'female': 0.08, 'other': 0.08},
 9: {'male': 0.09, 'female': 0.09, 'other': 0.09},
 10: {'male': 0.1, 'female': 0.1, 'other': 0.1},
 11: {'male': 0.11, 'female': 0.11, 'other': 0.11},
 12: {'male': 0.12, 'female': 0.12, 'other': 0.12},
 13: {'male': 0.13, 'female': 0.13, 'other': 0.13},
 14: {'male': 0.14, 'female': 0.14, 'other': 0.14}}

In [14]:

discrete_joint_distribution_sampler(Bobby, vulnerable_mapping, vulnerable_distribution)

discrete_joint_distribution_sampler(Bobby, vulnerable_mapping, vulnerable_distribution)

Out[14]:

0.06

Household Quarantine¶

Probabilistically apply quarantine to a household (remove all activities - stay at home)

Person-based, all people equal¶

If you have a probability of any person having to be quarantined

In [15]:

policy_quarantine = HouseholdPolicy(
    RemoveActivity(["work", "health", "leisure", "escort", "shop", "education"]),
    PersonProbability(0.0000000000001),
)

policy_quarantine = HouseholdPolicy( RemoveActivity(["work", "health", "leisure", "escort", "shop", "education"]), PersonProbability(0.0000000000001), )

In [16]:

q_pop = apply_policies(population, [policy_quarantine])
print_simple_stats(q_pop)

q_pop = apply_policies(population, [policy_quarantine]) print_simple_stats(q_pop)

Population total time at home: 74.58 hours
Population total travel time: 9.67 hours

In [17]:

q_pop.households["Smith"].plot()

q_pop.households["Smith"].plot()

In [18]:

q_pop.households["Jones"].plot()

q_pop.households["Jones"].plot()

Using joint distribution¶

In [19]:

policy_quarantine = HouseholdPolicy(
    RemoveActivity(["work", "health", "escort", "leisure", "shop", "education"]),
    PersonProbability(
        discrete_joint_distribution_sampler,
        {"mapping": vulnerable_mapping, "distribution": vulnerable_distribution},
    ),
)

policy_quarantine = HouseholdPolicy( RemoveActivity(["work", "health", "escort", "leisure", "shop", "education"]), PersonProbability( discrete_joint_distribution_sampler, {"mapping": vulnerable_mapping, "distribution": vulnerable_distribution}, ), )

In [20]:

q_pop = apply_policies(population, [policy_quarantine])
print_simple_stats(q_pop)

q_pop = apply_policies(population, [policy_quarantine]) print_simple_stats(q_pop)

Population total time at home: 41.42 hours
Population total travel time: 5.92 hours

In [21]:

q_pop.households["Jones"].plot()

q_pop.households["Jones"].plot()

Chaining probabilities¶

In [22]:

policy_quarantine = HouseholdPolicy(
    RemoveActivity(["work", "health", "escort", "leisure", "shop", "education"]),
    [
        PersonProbability(0.5),
        PersonProbability(
            discrete_joint_distribution_sampler,
            {"mapping": vulnerable_mapping, "distribution": vulnerable_distribution},
        ),
    ],
)

policy_quarantine = HouseholdPolicy( RemoveActivity(["work", "health", "escort", "leisure", "shop", "education"]), [ PersonProbability(0.5), PersonProbability( discrete_joint_distribution_sampler, {"mapping": vulnerable_mapping, "distribution": vulnerable_distribution}, ), ], )

In [23]:

q_pop = apply_policies(population, [policy_quarantine])
print_simple_stats(q_pop)

q_pop = apply_policies(population, [policy_quarantine]) print_simple_stats(q_pop)

Population total time at home: 43.67 hours
Population total travel time: 1.67 hours

In [24]:

q_pop.households["Jones"].plot()

q_pop.households["Jones"].plot()

Remove Education¶

Probabilistically remove education activities from a person and escort from people in the same household

In [25]:

edu_mapping = ["job"]
edu_distribution = {"employed": 0, "unemployed": 0, "education": 1}

edu_mapping = ["job"] edu_distribution = {"employed": 0, "unemployed": 0, "education": 1}

In [26]:

key_mapping = ["key_worker"]
key_distribution = {True: 0, False: 1}

key_mapping = ["key_worker"] key_distribution = {True: 0, False: 1}

In [27]:

edu_distribution, key_distribution

edu_distribution, key_distribution

Out[27]:

({'employed': 0, 'unemployed': 0, 'education': 1}, {True: 0, False: 1})

In [28]:

policy_remove_education_and_escort = HouseholdPolicy(
    RemoveActivity(["education", "escort"]),
    [
        PersonProbability(
            discrete_joint_distribution_sampler,
            {"mapping": edu_mapping, "distribution": edu_distribution},
        ),
        PersonProbability(
            discrete_joint_distribution_sampler,
            {"mapping": key_mapping, "distribution": key_distribution},
        ),
    ],
)

policy_remove_education_and_escort = HouseholdPolicy( RemoveActivity(["education", "escort"]), [ PersonProbability( discrete_joint_distribution_sampler, {"mapping": edu_mapping, "distribution": edu_distribution}, ), PersonProbability( discrete_joint_distribution_sampler, {"mapping": key_mapping, "distribution": key_distribution}, ), ], )

In [29]:

edu_pop = apply_policies(population, [policy_remove_education_and_escort])
print_simple_stats(edu_pop)

edu_pop = apply_policies(population, [policy_remove_education_and_escort]) print_simple_stats(edu_pop)

Population total time at home: 84.25 hours
Population total travel time: 7.83 hours

In [30]:

edu_pop.households["Smith"].people["Bradly"].plan.print()

edu_pop.households["Smith"].people["Bradly"].plan.print()

0:	Activity(act:home, location:a, time:00:00:00 --> 08:00:00, duration:8:00:00)
1:	Leg(mode:walk, area:a --> b, time:08:00:00 --> 08:20:00, duration:0:20:00)
2:	Activity(act:work, location:b, time:08:20:00 --> 09:20:00, duration:1:00:00)
3:	Leg(mode:pt, area:b --> c, time:09:20:00 --> 09:40:00, duration:0:20:00)
4:	Activity(act:work, location:b, time:09:40:00 --> 11:20:00, duration:1:40:00)
5:	Leg(mode:pt, area:b --> c, time:11:20:00 --> 11:40:00, duration:0:20:00)
6:	Activity(act:work, location:b, time:11:40:00 --> 13:20:00, duration:1:40:00)
7:	Leg(mode:pt, area:b --> c, time:13:20:00 --> 13:40:00, duration:0:20:00)
8:	Activity(act:leisure, location:c, time:13:40:00 --> 14:50:00, duration:1:10:00)
9:	Leg(mode:pt, area:c --> a, time:14:50:00 --> 15:10:00, duration:0:20:00)
10:	Activity(act:home, location:a, time:15:10:00 --> 16:50:00, duration:1:40:00)
11:	Leg(mode:car, area:a --> b, time:16:50:00 --> 17:10:00, duration:0:20:00)
12:	Activity(act:shop_1, location:b, time:17:10:00 --> 17:40:00, duration:0:30:00)
13:	Leg(mode:car, area:b --> b, time:17:40:00 --> 17:50:00, duration:0:10:00)
14:	Activity(act:shop_2, location:b, time:17:50:00 --> 18:40:00, duration:0:50:00)
15:	Leg(mode:car, area:b --> a, time:18:40:00 --> 19:00:00, duration:0:20:00)
16:	Activity(act:home, location:a, time:19:00:00 --> 00:00:00, duration:5:00:00)

In [31]:

q_pop.households["Smith"].plot()

q_pop.households["Smith"].plot()

In [32]:

edu_pop.households["Smith"].plot()

edu_pop.households["Smith"].plot()

In [33]:

edu_pop.households["Jones"].plot()

edu_pop.households["Jones"].plot()

In [34]:

edu_pop.households["Jones"].people["Bridget"].plan.print()

edu_pop.households["Jones"].people["Bridget"].plan.print()

0:	Activity(act:home, location:a, time:00:00:00 --> 08:00:00, duration:8:00:00)
1:	Leg(mode:walk, area:a --> b, time:08:00:00 --> 08:05:00, duration:0:05:00)
2:	Activity(act:escort, location:b, time:08:05:00 --> 10:30:00, duration:2:25:00)
3:	Leg(mode:pt, area:b --> c, time:10:30:00 --> 11:00:00, duration:0:30:00)
4:	Activity(act:work, location:c, time:11:00:00 --> 16:00:00, duration:5:00:00)
5:	Leg(mode:pt, area:c --> a, time:16:00:00 --> 16:20:00, duration:0:20:00)
6:	Activity(act:home, location:a, time:16:20:00 --> 17:20:00, duration:1:00:00)
7:	Leg(mode:pt, area:c --> a, time:17:20:00 --> 17:50:00, duration:0:30:00)
8:	Activity(act:shop, location:a, time:17:50:00 --> 18:30:00, duration:0:40:00)
9:	Leg(mode:pt, area:c --> a, time:18:30:00 --> 18:50:00, duration:0:20:00)
10:	Activity(act:home, location:a, time:18:50:00 --> 00:00:00, duration:5:10:00)

PersonAttributeFilter¶

You can also use the modify.PersonAttributeFilter to only affect people with certain attributes.

In [35]:

def condition_job_education(val):
    return val == "education"


policy_remove_education_and_escort = HouseholdPolicy(
    RemoveActivity(["education", "escort"]),
    PersonProbability(
        discrete_joint_distribution_sampler,
        {"mapping": key_mapping, "distribution": key_distribution},
    ),
    PersonAttributeFilter({"job": condition_job_education}),
)

def condition_job_education(val): return val == "education" policy_remove_education_and_escort = HouseholdPolicy( RemoveActivity(["education", "escort"]), PersonProbability( discrete_joint_distribution_sampler, {"mapping": key_mapping, "distribution": key_distribution}, ), PersonAttributeFilter({"job": condition_job_education}), )

In [36]:

apply_policies(population, [policy_remove_education_and_escort]).households["Smith"].plot()

apply_policies(population, [policy_remove_education_and_escort]).households["Smith"].plot()

In [37]:

policy_remove_education_and_escort = HouseholdPolicy(
    RemoveActivity(["education", "escort"]),
    [
        ActivityProbability(["education"], 0.9999),
        PersonProbability(
            discrete_joint_distribution_sampler,
            {"mapping": edu_mapping, "distribution": edu_distribution},
        ),
        PersonProbability(
            discrete_joint_distribution_sampler,
            {"mapping": key_mapping, "distribution": key_distribution},
        ),
    ],
)

policy_remove_education_and_escort = HouseholdPolicy( RemoveActivity(["education", "escort"]), [ ActivityProbability(["education"], 0.9999), PersonProbability( discrete_joint_distribution_sampler, {"mapping": edu_mapping, "distribution": edu_distribution}, ), PersonProbability( discrete_joint_distribution_sampler, {"mapping": key_mapping, "distribution": key_distribution}, ), ], )

In [38]:

apply_policies(population, [policy_remove_education_and_escort]).households["Smith"].plot()

apply_policies(population, [policy_remove_education_and_escort]).households["Smith"].plot()

Remove Leisure Activities¶

Remove all leisure activities

In [39]:

policy_remove_leisure = PersonPolicy(RemoveActivity(["leisure"]), PersonProbability(1.0))

policy_remove_leisure = PersonPolicy(RemoveActivity(["leisure"]), PersonProbability(1.0))

In [40]:

lei_pop = apply_policies(population, [policy_remove_leisure])
print_simple_stats(lei_pop)

lei_pop = apply_policies(population, [policy_remove_leisure]) print_simple_stats(lei_pop)

Population total time at home: 76.08 hours
Population total travel time: 9.33 hours

In [41]:

lei_pop.households["Smith"].plot()

lei_pop.households["Smith"].plot()

Unemployment/Furlough¶

Probabilistically remove all work activities from a person

In [42]:

key_mapping = ["key_worker"]
key_distribution = {True: 0, False: 1}

key_mapping = ["key_worker"] key_distribution = {True: 0, False: 1}

In [43]:

policy_unemployment_and_furlough = PersonPolicy(
    RemoveActivity(["work"]),
    [
        PersonProbability(
            discrete_joint_distribution_sampler,
            {"mapping": key_mapping, "distribution": key_distribution},
        )
    ],
)

policy_unemployment_and_furlough = PersonPolicy( RemoveActivity(["work"]), [ PersonProbability( discrete_joint_distribution_sampler, {"mapping": key_mapping, "distribution": key_distribution}, ) ], )

In [44]:

fur_pop = apply_policies(population, [policy_unemployment_and_furlough])
print_simple_stats(fur_pop)

fur_pop = apply_policies(population, [policy_unemployment_and_furlough]) print_simple_stats(fur_pop)

Population total time at home: 65.33 hours
Population total travel time: 8.58 hours

In [45]:

fur_pop.households["Smith"].plot()

fur_pop.households["Smith"].plot()

In [46]:

fur_pop.households["Jones"].plot()

fur_pop.households["Jones"].plot()

In [47]:

policy_reduced_work = ActivityPolicy(RemoveActivity(["work"]), ActivityProbability(["work"], 0.5))

apply_policies(population, [policy_reduced_work]).households["Smith"].plot()

policy_reduced_work = ActivityPolicy(RemoveActivity(["work"]), ActivityProbability(["work"], 0.5)) apply_policies(population, [policy_reduced_work]).households["Smith"].plot()

Remove Shopping¶

Probabilistically remove individual shopping activities from a person

In [48]:

policy_remove_shopping = PersonPolicy(RemoveActivity(["shop"]), ActivityProbability(["shop"], 1.0))

policy_remove_shopping = PersonPolicy(RemoveActivity(["shop"]), ActivityProbability(["shop"], 1.0))

In [49]:

shop_pop = apply_policies(population, [policy_remove_shopping])
print_simple_stats(shop_pop)

shop_pop = apply_policies(population, [policy_remove_shopping]) print_simple_stats(shop_pop)

Population total time at home: 76.08 hours
Population total travel time: 8.83 hours

In [50]:

shop_pop.households["Jones"].people["Bridget"].plot()

shop_pop.households["Jones"].people["Bridget"].plot()

Reducing shared activities¶

In [51]:

policy_reduce_shopping_activities = HouseholdPolicy(
    ReduceSharedActivity(["shop_1", "shop_2", "shop"]),
    ActivityProbability(["shop_1", "shop_2", "shop"], 1.0),
)

policy_reduce_shopping_activities = HouseholdPolicy( ReduceSharedActivity(["shop_1", "shop_2", "shop"]), ActivityProbability(["shop_1", "shop_2", "shop"], 1.0), )

In [52]:

population["Smith"].shared_activities()

population["Smith"].shared_activities()

Out[52]:

[<pam.activity.Activity at 0x137e48670>,
 <pam.activity.Activity at 0x137e359c0>,
 <pam.activity.Activity at 0x137e36170>,
 <pam.activity.Activity at 0x137e48d30>,
 <pam.activity.Activity at 0x137e48c70>]

In [53]:

shop_reduce_pop = apply_policies(population, [policy_reduce_shopping_activities])
print_simple_stats(shop_reduce_pop)

shop_reduce_pop = apply_policies(population, [policy_reduce_shopping_activities]) print_simple_stats(shop_reduce_pop)

Population total time at home: 76.75 hours
Population total travel time: 8.83 hours

In [54]:

shop_reduce_pop.households["Smith"].plot()

shop_reduce_pop.households["Smith"].plot()

In [55]:

shop_reduce_pop.households["Smith"].people["Bobby"].print()

shop_reduce_pop.households["Smith"].people["Bobby"].print()

Person: Bobby
{'age': 6, 'job': 'education', 'gender': 'male', 'key_worker': False}
0:	Activity(act:home, location:a, time:00:00:00 --> 08:00:00, duration:8:00:00)
1:	Leg(mode:walk, area:a --> b, time:08:00:00 --> 08:25:00, duration:0:25:00)
2:	Activity(act:education, location:b, time:08:25:00 --> 15:55:00, duration:7:30:00)
3:	Leg(mode:walk, area:b --> c, time:15:55:00 --> 16:30:00, duration:0:35:00)
4:	Activity(act:home, location:a, time:16:30:00 --> 00:00:00, duration:7:30:00)

Moving Shopping tours¶

In [56]:

policy_move_shopping_tours = PersonPolicy(
    MoveActivityTourToHomeLocation(["shop_1", "shop_2"]),
    [ActivityProbability(["shop_1", "shop_2"], 1.0)],
)

policy_move_shopping_tours = PersonPolicy( MoveActivityTourToHomeLocation(["shop_1", "shop_2"]), [ActivityProbability(["shop_1", "shop_2"], 1.0)], )

In [57]:

shop_tour_pop = apply_policies(population, [policy_move_shopping_tours])
print_simple_stats(shop_tour_pop)

shop_tour_pop = apply_policies(population, [policy_move_shopping_tours]) print_simple_stats(shop_tour_pop)

Population total time at home: 74.58 hours
Population total travel time: 9.67 hours

In [58]:

shop_tour_pop.households["Smith"].plot()

shop_tour_pop.households["Smith"].plot()

In [59]:

# above is equivalent to
policy_move_shopping_tours = PersonPolicy(
    MoveActivityTourToHomeLocation(["shop_1", "shop_2"]),
    [ActivityProbability(["shop_1"], 1.0), ActivityProbability(["shop_2"], 1.0)],
)

# above is equivalent to policy_move_shopping_tours = PersonPolicy( MoveActivityTourToHomeLocation(["shop_1", "shop_2"]), [ActivityProbability(["shop_1"], 1.0), ActivityProbability(["shop_2"], 1.0)], )

In [60]:

shop_tour_pop = apply_policies(population, [policy_move_shopping_tours])
print_simple_stats(shop_tour_pop)

shop_tour_pop = apply_policies(population, [policy_move_shopping_tours]) print_simple_stats(shop_tour_pop)

Population total time at home: 74.58 hours
Population total travel time: 9.67 hours

In [61]:

shop_tour_pop.households["Smith"].plot()

shop_tour_pop.households["Smith"].plot()

In [62]:

shop_tour_pop.households["Smith"].people["Bradly"].print()

shop_tour_pop.households["Smith"].people["Bradly"].print()

Person: Bradly
{'age': 40, 'job': 'employed', 'gender': 'male', 'key_worker': True}
0:	Activity(act:home, location:a, time:00:00:00 --> 08:00:00, duration:8:00:00)
1:	Leg(mode:walk, area:a --> b, time:08:00:00 --> 08:20:00, duration:0:20:00)
2:	Activity(act:escort, location:b, time:08:20:00 --> 08:30:00, duration:0:10:00)
3:	Leg(mode:pt, area:b --> b, time:08:30:00 --> 09:00:00, duration:0:30:00)
4:	Activity(act:work, location:b, time:09:00:00 --> 10:00:00, duration:1:00:00)
5:	Leg(mode:pt, area:b --> c, time:10:00:00 --> 10:20:00, duration:0:20:00)
6:	Activity(act:work, location:b, time:10:20:00 --> 12:00:00, duration:1:40:00)
7:	Leg(mode:pt, area:b --> c, time:12:00:00 --> 12:20:00, duration:0:20:00)
8:	Activity(act:work, location:b, time:12:20:00 --> 14:00:00, duration:1:40:00)
9:	Leg(mode:pt, area:b --> c, time:14:00:00 --> 14:20:00, duration:0:20:00)
10:	Activity(act:leisure, location:c, time:14:20:00 --> 15:30:00, duration:1:10:00)
11:	Leg(mode:pt, area:c --> b, time:15:30:00 --> 15:50:00, duration:0:20:00)
12:	Activity(act:escort, location:b, time:15:50:00 --> 16:00:00, duration:0:10:00)
13:	Leg(mode:walk, area:b --> a, time:16:00:00 --> 16:20:00, duration:0:20:00)
14:	Activity(act:home, location:a, time:16:20:00 --> 18:00:00, duration:1:40:00)
15:	Leg(mode:walk, area:a --> a, time:18:00:00 --> 18:20:00, duration:0:20:00)
16:	Activity(act:shop_1, location:a, time:18:20:00 --> 18:50:00, duration:0:30:00)
17:	Leg(mode:walk, area:a --> a, time:18:50:00 --> 19:00:00, duration:0:10:00)
18:	Activity(act:shop_2, location:a, time:19:00:00 --> 19:50:00, duration:0:50:00)
19:	Leg(mode:walk, area:a --> a, time:19:50:00 --> 20:10:00, duration:0:20:00)
20:	Activity(act:home, location:a, time:20:10:00 --> 00:00:00, duration:3:50:00)

All together now!¶

In [63]:

all_together_pop = apply_policies(
    population,
    [
        policy_quarantine,
        policy_remove_education_and_escort,
        policy_remove_leisure,
        policy_unemployment_and_furlough,
        policy_remove_shopping,
    ],
)

print_simple_stats(all_together_pop)

all_together_pop = apply_policies( population, [ policy_quarantine, policy_remove_education_and_escort, policy_remove_leisure, policy_unemployment_and_furlough, policy_remove_shopping, ], ) print_simple_stats(all_together_pop)

Population total time at home: 83.67 hours
Population total travel time: 4.25 hours