Synthetic Population Generation¶

Generate sythetic schedules and attributes.

In [1]:

from pathlib import Path

import pandas as pd

from caveat.data.synth import ActivityGen
from caveat.data.utils import generate_population_conditional, trace_to_pam
from caveat.evaluate.describe.times import (
    joint_time_distributions_plot,
    times_distributions_plot,
)
from caveat.evaluate.describe.transitions import sequence_prob_plot

from pathlib import Path import pandas as pd from caveat.data.synth import ActivityGen from caveat.data.utils import generate_population_conditional, trace_to_pam from caveat.evaluate.describe.times import ( joint_time_distributions_plot, times_distributions_plot, ) from caveat.evaluate.describe.transitions import sequence_prob_plot

In [2]:

write_path = Path("tmp")
n = 1000

write_path = Path("tmp") n = 1000

In [3]:

# FT worker generator
ftw = ActivityGen()
config = ftw.transition_config.copy()
for _, kv in config.items():
    for k, v in kv.items():
        if k in ["work"]:
            for i, (a, b) in enumerate(v):
                v[i] = (a, b * 10)
        if k in ["education", "shop", "leisure"]:
            for i, (a, b) in enumerate(v):
                v[i] = (a, b / 10)
ftw.build(config)

trace = ftw.run()
plan = trace_to_pam(trace, ftw.map)
plan.plot()

# FT worker generator ftw = ActivityGen() config = ftw.transition_config.copy() for _, kv in config.items(): for k, v in kv.items(): if k in ["work"]: for i, (a, b) in enumerate(v): v[i] = (a, b * 10) if k in ["education", "shop", "leisure"]: for i, (a, b) in enumerate(v): v[i] = (a, b / 10) ftw.build(config) trace = ftw.run() plan = trace_to_pam(trace, ftw.map) plan.plot()

In [4]:

# PT worker generator
ptw = ActivityGen()
config = ptw.transition_config.copy()
for o, kv in config.items():
    for k, v in kv.items():
        if not o == "work":
            continue
        if k == "work":
            for i, (a, b) in enumerate(v):
                v[i] = (a, b / 100)
        if k in ["shop", "leisure"]:
            for i, (a, b) in enumerate(v):
                v[i] = (a, b * 100)
ptw.build(config)

trace = ptw.run()
plan = trace_to_pam(trace, ptw.map)
plan.plot()

# PT worker generator ptw = ActivityGen() config = ptw.transition_config.copy() for o, kv in config.items(): for k, v in kv.items(): if not o == "work": continue if k == "work": for i, (a, b) in enumerate(v): v[i] = (a, b / 100) if k in ["shop", "leisure"]: for i, (a, b) in enumerate(v): v[i] = (a, b * 100) ptw.build(config) trace = ptw.run() plan = trace_to_pam(trace, ptw.map) plan.plot()

In [5]:

# Leisure generator
neet = ActivityGen()
config = neet.transition_config.copy()
for _, kv in config.items():
    for k, v in kv.items():
        if k in ["work", "education"]:
            for i, (a, b) in enumerate(v):
                v[i] = (a, 0)
        if k in ["shop", "leisure"]:
            for i, (a, b) in enumerate(v):
                v[i] = (a, b * 10)

neet.build(config)

trace = neet.run()
plan = trace_to_pam(trace, neet.map)
plan.plot()

# Leisure generator neet = ActivityGen() config = neet.transition_config.copy() for _, kv in config.items(): for k, v in kv.items(): if k in ["work", "education"]: for i, (a, b) in enumerate(v): v[i] = (a, 0) if k in ["shop", "leisure"]: for i, (a, b) in enumerate(v): v[i] = (a, b * 10) neet.build(config) trace = neet.run() plan = trace_to_pam(trace, neet.map) plan.plot()

In [6]:

# FTE generator
fte = ActivityGen()
config = fte.transition_config.copy()
for _, kv in config.items():
    for k, v in kv.items():
        if k == "work":
            for i, (a, b) in enumerate(v):
                v[i] = (a, 0)
        if k == "education":
            for i, (a, b) in enumerate(v):
                v[i] = (a, b * 100)

fte.build(config)

trace = fte.run()
plan = trace_to_pam(trace, fte.map)
plan.plot()

# FTE generator fte = ActivityGen() config = fte.transition_config.copy() for _, kv in config.items(): for k, v in kv.items(): if k == "work": for i, (a, b) in enumerate(v): v[i] = (a, 0) if k == "education": for i, (a, b) in enumerate(v): v[i] = (a, b * 100) fte.build(config) trace = fte.run() plan = trace_to_pam(trace, fte.map) plan.plot()

In [7]:

population = generate_population_conditional(gens=(ftw, ptw, neet, fte), size=n)
population.act = population.act.map(ftw.map)
population = population[
    ["pid", "act", "start", "end", "duration", "gender", "age", "employment"]
]
population

population = generate_population_conditional(gens=(ftw, ptw, neet, fte), size=n) population.act = population.act.map(ftw.map) population = population[ ["pid", "act", "start", "end", "duration", "gender", "age", "employment"] ] population

Out[7]:

	pid	act	start	end	duration	gender	age	employment
0	0	home	0	375	375	F	24	FTW
1	0	work	375	1170	795	F	24	FTW
2	0	home	1170	1440	270	F	24	FTW
3	1	home	0	390	390	M	85	NEET
4	1	shop	390	450	60	M	85	NEET
...	...	...	...	...	...	...	...	...
4305	999	home	0	390	390	F	83	NEET
4306	999	shop	390	450	60	F	83	NEET
4307	999	home	450	630	180	F	83	NEET
4308	999	leisure	630	990	360	F	83	NEET
4309	999	home	990	1440	450	F	83	NEET

4310 rows × 8 columns

In [8]:

# population.to_csv(write_path / "combined.csv", index=False)
population[["pid", "act", "start", "end", "duration"]].to_csv(
    write_path / "synthetic_schedules.csv", index=False
)
population[["pid", "gender", "age", "employment"]].drop_duplicates().to_csv(
    write_path / "synthetic_attributes.csv", index=False
)

# population.to_csv(write_path / "combined.csv", index=False) population[["pid", "act", "start", "end", "duration"]].to_csv( write_path / "synthetic_schedules.csv", index=False ) population[["pid", "gender", "age", "employment"]].drop_duplicates().to_csv( write_path / "synthetic_attributes.csv", index=False )

In [9]:

cs = population.gender.map({"M": "blue", "F": "red"})
population.plot(kind="scatter", x="age", y="duration", alpha=0.1, c=cs)

cs = population.gender.map({"M": "blue", "F": "red"}) population.plot(kind="scatter", x="age", y="duration", alpha=0.1, c=cs)

Out[9]:

<Axes: xlabel='age', ylabel='duration'>

In [10]:

def describe_col(population, col: str) -> pd.DataFrame:
    description = population.groupby("act")[col].describe()[
        ["count", "mean", "std", "min", "max"]
    ]
    description["attribute"] = col
    return description


def describe_cols(population, cols: list[str]) -> pd.DataFrame:
    description = pd.concat(
        [describe_col(population, c) for c in cols], ignore_index=False
    )
    description = description.reset_index().set_index(["attribute", "act"])
    return description


describe_cols(population, ["start", "end", "duration"]).round()

def describe_col(population, col: str) -> pd.DataFrame: description = population.groupby("act")[col].describe()[ ["count", "mean", "std", "min", "max"] ] description["attribute"] = col return description def describe_cols(population, cols: list[str]) -> pd.DataFrame: description = pd.concat( [describe_col(population, c) for c in cols], ignore_index=False ) description = description.reset_index().set_index(["attribute", "act"]) return description describe_cols(population, ["start", "end", "duration"]).round()

Out[10]:

		count	mean	std	min	max
attribute	act
start	education	34.0	960.0	129.0	765.0	1125.0
	home	2085.0	494.0	490.0	0.0	1290.0
	leisure	798.0	565.0	240.0	375.0	1260.0
	shop	896.0	529.0	188.0	375.0	1170.0
	work	497.0	570.0	299.0	375.0	1275.0
end	education	34.0	1038.0	74.0	885.0	1140.0
	home	2085.0	907.0	518.0	375.0	1440.0
	leisure	798.0	898.0	207.0	420.0	1275.0
	shop	896.0	623.0	193.0	390.0	1185.0
	work	497.0	1023.0	157.0	735.0	1290.0
duration	education	34.0	78.0	60.0	15.0	195.0
	home	2085.0	413.0	112.0	15.0	885.0
	leisure	798.0	333.0	205.0	15.0	630.0
	shop	896.0	94.0	70.0	15.0	420.0
	work	497.0	453.0	260.0	15.0	825.0

In [11]:

subpops = {
    "males": population.loc[population.gender == "M"],
    "females": population.loc[population.gender == "F"],
    "young": population.loc[population.age < 25],
    "middle": population.loc[(population.age >= 25) & (population.age < 65)],
    "old": population.loc[population.age >= 65],
}

subpops = { "males": population.loc[population.gender == "M"], "females": population.loc[population.gender == "F"], "young": population.loc[population.age < 25], "middle": population.loc[(population.age >= 25) & (population.age < 65)], "old": population.loc[population.age >= 65], }

In [12]:

_ = times_distributions_plot(population, ys=subpops)

_ = times_distributions_plot(population, ys=subpops)

In [13]:

_ = joint_time_distributions_plot(population, ys=subpops, figsize=(12, 12))

_ = joint_time_distributions_plot(population, ys=subpops, figsize=(12, 12))

In [14]:

_ = sequence_prob_plot(population, ys=subpops, figsize=(12, 4))

_ = sequence_prob_plot(population, ys=subpops, figsize=(12, 4))

In [ ]: