def create_slider(districts_file, population_file, households_file, persons_file, transmission_log):
"""
Create plot with interactive slider.
"""
#######################################
# Gather information about districts. #
#######################################
districts = {}
districts[0] = (0,0,0)
latitudes = []
longitudes = []
with open(districts_file, 'r') as district_file:
reader = csv.DictReader(district_file, delimiter=';')
for row in reader:
d_id = int(row['id'])
x = float(row['latitude'].replace(",", "."))
y = float(row['longitude'].replace(",", "."))
pop_size = int(row['pop_size'])
# Don't add districts with incorrect location data
if abs(x) < 90 and abs(y) < 180:
latitudes.append(x)
longitudes.append(y)
districts[d_id] = (x, y, pop_size)
# Use Smopy to get map image for the given region (calculated from min/max latitudes/longitudes in districts file)
region_map = smopy.Map((min(latitudes), min(longitudes), max(latitudes), max(longitudes)), z=9)
##################################################################
# Gather information about persons and households in simulation. #
##################################################################
# Check how long each person stays ill (= infectious) when infected.
person_infectiousness = {}
with open(persons_file, 'r') as person_file:
reader = csv.DictReader(person_file, delimiter=';')
for row in reader:
person_infectiousness[int(row['id'])] = row['end_infectiousness']
# For each person, fetch the household ID
person_households = []
with open (population_file, 'r') as pop_file:
reader = csv.DictReader(pop_file, delimiter=";")
for row in reader:
person_households.append(int(float(row['household_id'])))
# For each household, fetch the district it belongs to.
households = {}
with open(households_file, 'r') as hh_file:
reader = csv.DictReader(hh_file, delimiter=';')
for row in reader:
hh_id = int(row['id'])
d_id = int(row['district_id'])
households[hh_id] = d_id
########################################
# Gather information about infections. #
########################################
prepare_csv_transmissions(transmission_log, "infections.csv")
infections = {}
with open('./infections.csv', 'r') as infections_file:
reader = csv.DictReader(infections_file, delimiter=",")
for row in reader:
person_id = int(row['person_id'])
if (person_id < len(person_infectiousness) + 1):
hh_id = person_households[person_id]
district = households[hh_id]
start_infected = int(row['begin_infection'])
end_infected = start_infected + int(person_infectiousness[person_id])
for day in range(start_infected, end_infected + 1):
if day in infections:
# add extra infection
if district in infections[day]:
infections[day][district] += 1
else:
infections[day][district] = 1
else:
# add first infection
infections[day] = {}
infections[day][district] = 1
##################
# Create figure. #
##################
fig, ax = plt.subplots()
fig.canvas.set_window_title("Evolution of disease spread")
plt.subplots_adjust(bottom=0.25)
# Legend for circles sizes (representing district population sizes).
l1 = plt.scatter([],[], s=4, facecolors='none', edgecolors='black')
l2 = plt.scatter([],[], s=16, facecolors='none', edgecolors='black')
l3 = plt.scatter([],[], s=64, facecolors='none', edgecolors='black')
l4 = plt.scatter([],[], s=256, facecolors='none', edgecolors='black')
l5 = plt.scatter([],[], s=1024, facecolors='none', edgecolors='black')
labels = ["< 224", "< 327", "< 525", "< 5000", "> 5000"]
leg = plt.legend([l1, l2, l3, l4, l5], labels, ncol=5, frameon=False, fontsize=10,
handlelength=2, loc = 8, bbox_to_anchor = [0.5,-0.15],
handletextpad=1, title='District sizes', scatterpoints = 1)
leg.get_title().set_position((0, -35))
# Scatterplot to represent districts + fraction of infected per district.
region_map.show_mpl(ax)
l = plt.scatter([], [], alpha=0.5)
l.axes.get_xaxis().set_ticks([])
l.axes.get_yaxis().set_ticks([])
axcolor = 'lightgoldenrodyellow'
ax_day = plt.axes([0.12, 0.1, 0.79, 0.03], axisbg=axcolor)
sday = Slider(ax_day, 'Day', 0.1, 100.0, valinit=1, valfmt='%0.0f')
plt.subplots_adjust(bottom=0.15)
# Legend for colors (representing fraction of infected in district).
data_range = [0, 100]
m = cm.ScalarMappable(cmap=cm.autumn_r)
m.set_array(data_range)
c = plt.colorbar(m,ax=ax, shrink=0.5, orientation="horizontal")
c.set_label("Fraction of district infected")
###############################
# Update function for slider. #
###############################
def update(val):
"""
Update the plot according to the value of the slider.
"""
day = int(round(sday.val))
xs = []
ys = []
offsets = []
sizes = []
colors = []
if (day in infections):
day_infections = infections[day]
for key, value in districts.iteritems():
x, y = region_map.to_pixels((value[0], value[1]))
district_size = value[2]
circle_size = 0
if (district_size <= 224):
circle_size = 4
elif (district_size <= 327):
circle_size = 16
elif (district_size <= 525):
circle_size = 64
elif (district_size <= 5000):
circle_size = 256
else:
circle_size = 1024
if key in day_infections:
num_infections = day_infections[key]
xs.append(x)
ys.append(y)
offsets.append([x, y])
sizes.append(circle_size)
colors.append([1, 1 - (float(num_infections) / district_size), 0])
l.set_offsets(offsets)
l.set_sizes(sizes)
l.set_facecolors(np.array(colors))
fig.canvas.draw_idle()
# Execute update() when value of slider changes
sday.on_changed(update)
plt.show()
def create_map(district_file, pop_file, households_file, persons_file,
transmissions_log):
"""
Create .png map image for each simulation day in the given files.
"""
########################################################################
# Gather information (location & population size) about the districts. #
########################################################################
districts = {}
districts[0] = (0, 0, 0)
latitudes = []
longitudes = []
with open(district_file, 'r') as district_file:
reader = csv.DictReader(district_file, delimiter=";")
for row in reader:
d_id = int(row['id'])
latitude = float(row['latitude'].replace(",", "."))
longitude = float(row['longitude'].replace(",", "."))
pop_size = int(row['pop_size'])
# Don't add districts with incorrect location data
if abs(latitude) < 90 and abs(longitude) < 180:
districts[d_id] = (latitude, longitude, pop_size)
latitudes.append(latitude)
longitudes.append(longitude)
# Use Smopy to get map image for the given region (calculated from min/max latitudes/longitudes in districts file)
region_map = smopy.Map(
(min(latitudes), min(longitudes), max(latitudes), max(longitudes)),
z=9)
##################################################################
# Gather information about persons and households in simulation. #
##################################################################
# Check how long each person stays ill (= infectious) when infected.
person_infectiousness = {}
with open(persons_file, 'r') as person_file:
reader = csv.DictReader(person_file, delimiter=';')
for row in reader:
person_infectiousness[int(row['id'])] = row['end_infectiousness']
# For each person, fetch the household ID
person_households = []
with open(pop_file, 'r') as pop_file:
reader = csv.DictReader(pop_file, delimiter=";")
for row in reader:
person_households.append(int(float(row['household_id'])))
# For each household, fetch the district it belongs to.
households = {}
with open(households_file, 'r') as hh_file:
reader = csv.DictReader(hh_file, delimiter=';')
for row in reader:
hh_id = int(row['id'])
d_id = int(row['district_id'])
households[hh_id] = d_id
########################################
# Gather information about infections. #
########################################
prepare_csv_transmissions(transmissions_log, "infections.csv")
infections = {}
max_infect = 0
with open('./infections.csv', 'r') as infections_file:
reader = csv.DictReader(infections_file, delimiter=",")
for row in reader:
person_id = int(row['person_id'])
if (person_id < len(person_infectiousness) + 1):
hh_id = person_households[person_id]
district = households[hh_id]
start_infected = int(row['begin_infection'])
end_infected = start_infected + int(
person_infectiousness[person_id])
for day in range(start_infected, end_infected + 1):
if day in infections:
# add extra infection
if district in infections[day]:
infections[day][district] += 1
if infections[day][district] > max_infect:
max_infect = infections[day][district]
else:
infections[day][district] = 1
else:
# add first infection
infections[day] = {}
infections[day][district] = 1
#########################################################
# Create the actual map images for each simulation day. #
#########################################################
for day in range(0, 100):
ax = region_map.show_mpl()
cases = False
if (day in infections):
cases = True
day_infections = infections[day]
for i, district_info in districts.iteritems():
x, y = region_map.to_pixels(district_info[0], district_info[1])
# Calculate circle size from district size
district_size = district_info[2]
circle_size = 0
if (district_size <= 224):
# first quartile
circle_size = 2
elif (district_size <= 327):
# second quartile
circle_size = 4
elif (district_size <= 525):
# third quartile
circle_size = 8
elif (district_size <= 5000):
# 4th quartile split in two, to accentuate really big districts
circle_size = 16
else:
circle_size = 32
# no infections -> districts are not shown
# otherwise: spectrum from red (few infections) to yellow (lot of infections)
if (cases) and (i in day_infections):
num_infections = day_infections[i]
R = 1
# normalize # of infections to number between 0 and 1
#G = float(num_infections) / max_infect
G = 1 - (float(num_infections) / district_size)
B = 0
col = (R, G, B)
ax.plot(x, y, marker='o', color=col, ms=circle_size, mew=1)
#############################################
# Create legends for circle sizes & colors. #
#############################################
l1 = plt.scatter([], [], s=4, facecolors='none', edgecolors='black')
l2 = plt.scatter([], [], s=16, facecolors='none', edgecolors='black')
l3 = plt.scatter([], [], s=64, facecolors='none', edgecolors='black')
l4 = plt.scatter([], [], s=256, facecolors='none', edgecolors='black')
l5 = plt.scatter([], [], s=1024, facecolors='none', edgecolors='black')
labels = ["< 224", "< 327", "< 525", "< 5000", "> 5000"]
leg = plt.legend([l1, l2, l3, l4, l5],
labels,
ncol=5,
frameon=False,
fontsize=10,
handlelength=2,
loc=8,
handletextpad=1,
title='District sizes',
scatterpoints=1,
bbox_to_anchor=(0.5, -0.15))
leg.get_title().set_position((0, -50))
data_range = [0, 100]
m = cm.ScalarMappable(cmap=cm.autumn_r)
m.set_array(data_range)
c = plt.colorbar(m, orientation="horizontal")
c.set_label("Fraction of district infected")
###########################
# Save map image to .png. #
###########################
fig_title = "day " + str(day)
plt.title(fig_title)
fig_file = "map_" + str(day) + ".png"
plt.savefig(fig_file)