def __init__(self, borders_file, output_file, region, photo_dens_file=None, pop_dens_file=None,
top_cities_file=None, osm=False, resolution='i', width=50., thick=10., color='darkred',
linestyle='solid', font_size=30.0, dot_size=30.0, label_offset=0.00075, sep=1.0, intervals=100,
scale_sizes='', natural_scales=''):
self.sep = sep
self.intervals = intervals
if len(scale_sizes) == 0:
self.scale_sizes = []
else:
self.scale_sizes = [float(token) for token in scale_sizes.split(',')]
if len(natural_scales) == 0:
self.natural_scales = []
else:
self.natural_scales = [float(token) for token in natural_scales.split(',')]
extra_height = (len(self.scale_sizes) + 1) * 0.035
DrawMap.__init__(self, borders_file, output_file, region, photo_dens_file, pop_dens_file, top_cities_file, osm,
resolution, width, thick, color, linestyle, font_size, dot_size, label_offset, extra_height)
self.width_c = self.cc[3] - self.cc[1]
self.height_c = self.cc[2] - self.cc[0]
self.width_km = geo.distance({'lat': self.cc[0], 'lng': self.cc[1]},
{'lat': self.cc[0], 'lng': self.cc[3]})
self.height_km = geo.distance({'lat': self.cc[0], 'lng': self.cc[1]},
{'lat': self.cc[2], 'lng': self.cc[1]})
print('width: %s; height: %s' % (self.width_c, self.height_c))
print('width: %skm; height: %skm' % (self.width_km, self.height_km))
def tokms(segments):
min_x = float('inf')
min_y = float('inf')
for seg in segments:
if seg['x1'] < min_x:
min_x = seg['x1']
if seg['x2'] < min_x:
min_x = seg['x2']
if seg['y1'] < min_y:
min_y = seg['y1']
if seg['x2'] < min_x:
min_y = seg['y2']
normsegs = []
for seg in segments:
nseg = {
'x1':
geo.distance({
'lat': min_x,
'lng': min_y
}, {
'lat': seg['x1'],
'lng': min_y
}),
'y1':
geo.distance({
'lat': min_x,
'lng': min_y
}, {
'lat': min_x,
'lng': seg['y1']
}),
'x2':
geo.distance({
'lat': min_x,
'lng': min_y
}, {
'lat': seg['x2'],
'lng': min_y
}),
'y2':
geo.distance({
'lat': min_x,
'lng': min_y
}, {
'lat': min_x,
'lng': seg['y2']
}),
'id1':
seg['id1'],
'id2':
seg['id2']
}
normsegs.append(nseg)
return normsegs
def process_link(self, link, time):
v1 = link[0]
v2 = link[1]
if v1 > v2:
v1 = link[1]
v2 = link[0]
l = (v1, v2)
if l in self.ll:
self.ll[l] += 1
else:
self.ll[l] = 1
self.locmap[v1]['neighbors'] += 1
self.locmap[v2]['neighbors'] += 1
self.locmap[v1]['degree'] += 1
self.locmap[v2]['degree'] += 1
loc1 = self.locmap[v1]
loc2 = self.locmap[v2]
dist = geo.distance(loc1, loc2)
self.locmap[v1]['dist'] += dist
self.locmap[v2]['dist'] += dist
t = float(time) / 60.
t /= 60.
t /= 24.
self.locmap[v1]['time'] += t
self.locmap[v2]['time'] += t
self.update_angle(v1, v2)
self.update_angle(v2, v1)
def post_process(self):
for loc in self.locmap:
degree = self.locmap[loc]['degree']
if degree > 0:
self.locmap[loc]['dist'] /= degree
self.locmap[loc]['time'] /= degree
self.locmap[loc]['angle'] /= degree
self.locmap[loc]['users'] = len(self.locmap[loc]['users'])
for link in self.ll:
v1 = link[0]
v2 = link[1]
w = float(self.ll[link])
loc1 = self.locmap[v1]
loc2 = self.locmap[v2]
dist = geo.distance(loc1, loc2)
self.update_entropy(v1, w)
self.update_entropy(v2, w)
self.update_dist_var(v1, dist, w)
self.update_dist_var(v2, dist, w)
self.update_angle_var(v1, v2, w)
self.update_angle_var(v2, v1, w)
self.update_angle_entropy(v1, v2, w)
self.update_angle_entropy(v2, v1, w)
for loc in self.locmap:
degree = self.locmap[loc]['degree']
if degree > 0:
self.locmap[loc]['dist_var'] /= degree
self.locmap[loc]['angle_var'] /= degree
self.compute_angle_entropy(loc)
def filter(self, csv_in, csv_out, max_dist):
print('filtering graph for maximum distance: %s' % max_dist)
g = graph.read_graph(csv_in)
for edge in g:
loc1 = self.locmap.coords[edge[0]]
loc2 = self.locmap.coords[edge[1]]
dist = geo.distance(loc1, loc2)
if dist > max_dist:
g[edge] = 0.
graph.write_graph(g, csv_out)
def tokms(segments):
min_x = float('inf')
min_y = float('inf')
for seg in segments:
if seg['x1'] < min_x:
min_x = seg['x1']
if seg['x2'] < min_x:
min_x = seg['x2']
if seg['y1'] < min_y:
min_y = seg['y1']
if seg['x2'] < min_x:
min_y = seg['y2']
normsegs = []
for seg in segments:
nseg = {'x1': geo.distance({'lat': min_x, 'lng': min_y}, {'lat': seg['x1'], 'lng': min_y}),
'y1': geo.distance({'lat': min_x, 'lng': min_y}, {'lat': min_x, 'lng': seg['y1']}),
'x2': geo.distance({'lat': min_x, 'lng': min_y}, {'lat': seg['x2'], 'lng': min_y}),
'y2': geo.distance({'lat': min_x, 'lng': min_y}, {'lat': min_x, 'lng': seg['y2']}),
'id1': seg['id1'],
'id2': seg['id2']}
normsegs.append(nseg)
return normsegs
def write_dists(g, db, file_path):
f_dist = open(file_path, 'w')
f_dist.write('distance\n')
locmap = LocMap(db)
for edge in g:
loc1 = locmap.coords[edge[0]]
loc2 = locmap.coords[edge[1]]
dist = geo.distance(loc1, loc2)
if dist > 0:
for i in range(int(g[edge])):
f_dist.write('%s\n' % (dist,))
else:
print('zero distance found between %s and %s' % (loc1, loc2))
f_dist.close()
def compute(self, infile, outfile):
g = graph.read_graph(infile)
f = open(outfile, 'w')
total_distance = 0.
count = 0.
for edge in g:
loc1 = self.locmap.coords[edge[0]]
loc2 = self.locmap.coords[edge[1]]
weight = g[edge]
dist = geo.distance(loc1, loc2)
for i in range(weight):
f.write('%s\n' % dist)
total_distance += dist
count += 1.
f.close()
mean_distance = total_distance / count
print('mean distance: %s' % mean_distance)
def compute(self, infile, outfile):
g = graph.read_graph(infile)
f = open(outfile, "w")
total_distance = 0.0
count = 0.0
for edge in g:
loc1 = self.locmap.coords[edge[0]]
loc2 = self.locmap.coords[edge[1]]
weight = g[edge]
dist = geo.distance(loc1, loc2)
for i in range(weight):
f.write("%s\n" % dist)
total_distance += dist
count += 1.0
f.close()
mean_distance = total_distance / count
print("mean distance: %s" % mean_distance)
def process_user(self, user_id):
self.db.cur.execute("SELECT location, ts, id FROM media WHERE user=%s ORDER BY ts" % (user_id, ))
data = self.db.cur.fetchall()
locations = [x[0] for x in data]
# unique locations
ulocations = set(locations)
# only compute metrics for users who have been to at least 2 distinct locations
if len(ulocations) >= 2:
# photos
photos = len(data)
photo_ids = [x[2] for x in data]
# time stuff
times = [x[1] for x in data]
times.sort()
first_ts = min(times)
last_ts = max(times)
time_deltas = [times[i] - times[i - 1] for i in range(1, len(times))]
mean_time_interval = sum(t for t in time_deltas) / len(time_deltas)
# location stuff
loc_count = len(ulocations)
freqs = {}
for loc in ulocations:
freqs[loc] = 0
for loc in locations:
freqs[loc] += 1
herfindahl = 0.0
for loc in freqs:
s = freqs[loc] / len(locations)
herfindahl += s * s
links = itertools.combinations(ulocations, 2)
distances = [geo.distance(self.locs[link[0]], self.locs[link[1]]) for link in links]
mean_distance = sum(distances) / len(distances)
dists_str = ' '.join([str(d) for d in distances])
total_dist = 0.
count = 0
for i in range(1, len(locations)):
loc0 = locations[i - 1]
loc1 = locations[i]
if loc0 != loc1:
dist = geo.distance(self.locs[loc0], self.locs[loc1])
total_dist += dist
count += 1
mean_weighted_dist = total_dist / count
# comments
self.db.cur.execute("SELECT count(id) FROM comment WHERE user=%s" % (user_id,))
data = self.db.cur.fetchall()
comments_given = data[0][0]
comments_received = self.x_received('comment', photo_ids)
# likes
self.db.cur.execute("SELECT count(id) FROM likes WHERE user=%s" % (user_id,))
data = self.db.cur.fetchall()
likes_given = data[0][0]
likes_received = self.x_received('comment', photo_ids)
# print('photos: %s; first_ts: %s; last_ts: %s; mean_time_interval: %s; loc_count: %s; herfindahl: %s; mean_distance: %s; mean_weighted_distance: %s; comments_given: %s; comments_received: %s; likes_given: %s; likes_received: %s'
# % (photos, first_ts, last_ts, mean_time_interval, loc_count, herfindahl, mean_distance, mean_weighted_dist, comments_given, comments_received, likes_given, likes_received))
self.db.cur.execute('UPDATE user SET active=1,dists_str="%s",photos=%s,first_ts=%s,last_ts=%s,mean_time_interval=%s,locations=%s,herfindahl=%s,mean_dist=%s,mean_weighted_dist=%s,comments_given=%s,comments_received=%s,likes_given=%s,likes_received=%s WHERE id=%s'
% (dists_str, photos, first_ts, last_ts, mean_time_interval, loc_count, herfindahl, mean_distance, mean_weighted_dist, comments_given, comments_received, likes_given, likes_received, user_id,))
self.db.conn.commit()
def process_user(self, user_id):
self.db.cur.execute(
"SELECT location, ts, id FROM media WHERE user=%s ORDER BY ts" %
(user_id, ))
data = self.db.cur.fetchall()
locations = [x[0] for x in data]
# unique locations
ulocations = set(locations)
# only compute metrics for users who have been to at least 2 distinct locations
if len(ulocations) >= 2:
# photos
photos = len(data)
photo_ids = [x[2] for x in data]
# time stuff
times = [x[1] for x in data]
times.sort()
first_ts = min(times)
last_ts = max(times)
time_deltas = [
times[i] - times[i - 1] for i in range(1, len(times))
]
mean_time_interval = sum(t for t in time_deltas) / len(time_deltas)
# location stuff
loc_count = len(ulocations)
freqs = {}
for loc in ulocations:
freqs[loc] = 0
for loc in locations:
freqs[loc] += 1
herfindahl = 0.0
for loc in freqs:
s = freqs[loc] / len(locations)
herfindahl += s * s
links = itertools.combinations(ulocations, 2)
distances = [
geo.distance(self.locs[link[0]], self.locs[link[1]])
for link in links
]
mean_distance = sum(distances) / len(distances)
dists_str = ' '.join([str(d) for d in distances])
total_dist = 0.
count = 0
for i in range(1, len(locations)):
loc0 = locations[i - 1]
loc1 = locations[i]
if loc0 != loc1:
dist = geo.distance(self.locs[loc0], self.locs[loc1])
total_dist += dist
count += 1
mean_weighted_dist = total_dist / count
# comments
self.db.cur.execute("SELECT count(id) FROM comment WHERE user=%s" %
(user_id, ))
data = self.db.cur.fetchall()
comments_given = data[0][0]
comments_received = self.x_received('comment', photo_ids)
# likes
self.db.cur.execute("SELECT count(id) FROM likes WHERE user=%s" %
(user_id, ))
data = self.db.cur.fetchall()
likes_given = data[0][0]
likes_received = self.x_received('comment', photo_ids)
# print('photos: %s; first_ts: %s; last_ts: %s; mean_time_interval: %s; loc_count: %s; herfindahl: %s; mean_distance: %s; mean_weighted_distance: %s; comments_given: %s; comments_received: %s; likes_given: %s; likes_received: %s'
# % (photos, first_ts, last_ts, mean_time_interval, loc_count, herfindahl, mean_distance, mean_weighted_dist, comments_given, comments_received, likes_given, likes_received))
self.db.cur.execute(
'UPDATE user SET active=1,dists_str="%s",photos=%s,first_ts=%s,last_ts=%s,mean_time_interval=%s,locations=%s,herfindahl=%s,mean_dist=%s,mean_weighted_dist=%s,comments_given=%s,comments_received=%s,likes_given=%s,likes_received=%s WHERE id=%s'
% (
dists_str,
photos,
first_ts,
last_ts,
mean_time_interval,
loc_count,
herfindahl,
mean_distance,
mean_weighted_dist,
comments_given,
comments_received,
likes_given,
likes_received,
user_id,
))
self.db.conn.commit()
def __init__(self,
borders_file,
output_file,
region,
photo_dens_file=None,
pop_dens_file=None,
top_cities_file=None,
osm=False,
resolution='i',
width=50.,
thick=10.,
color='darkred',
linestyle='solid',
font_size=30.0,
dot_size=30.0,
label_offset=0.00075,
sep=1.0,
intervals=100,
scale_sizes='',
natural_scales=''):
self.sep = sep
self.intervals = intervals
if len(scale_sizes) == 0:
self.scale_sizes = []
else:
self.scale_sizes = [
float(token) for token in scale_sizes.split(',')
]
if len(natural_scales) == 0:
self.natural_scales = []
else:
self.natural_scales = [
float(token) for token in natural_scales.split(',')
]
extra_height = (len(self.scale_sizes) + 1) * 0.035
DrawMap.__init__(self, borders_file, output_file, region,
photo_dens_file, pop_dens_file, top_cities_file, osm,
resolution, width, thick, color, linestyle, font_size,
dot_size, label_offset, extra_height)
self.width_c = self.cc[3] - self.cc[1]
self.height_c = self.cc[2] - self.cc[0]
self.width_km = geo.distance({
'lat': self.cc[0],
'lng': self.cc[1]
}, {
'lat': self.cc[0],
'lng': self.cc[3]
})
self.height_km = geo.distance({
'lat': self.cc[0],
'lng': self.cc[1]
}, {
'lat': self.cc[2],
'lng': self.cc[1]
})
print('width: %s; height: %s' % (self.width_c, self.height_c))
print('width: %skm; height: %skm' % (self.width_km, self.height_km))