def annual_population(in_df, out_df, o_path):
in_df = calls_cleanup(in_df)
out_df = calls_cleanup(out_df)
df = in_df.append(out_df, sort=False)
df = df.groupby(['MOUHAFAZA', 'YEAR', 'KAZA'],
as_index=False).agg({'NB_CALLS': 'sum'})
df = df.apply(lambda row: get_population(df, row), axis=1)
save(df, o_path)
def structure(df):
df2=pd.DataFrame()
for year in np.unique(np.array(df['YEAR'])):
sub=df[df['YEAR'] == year]
for age in np.unique(np.array(sub['AGE_GROUP'])):
m=-1*sub[(sub['YEAR']==year)&(sub['AGE_GROUP']==age)&(sub['GENDER']=='M')]['NB_CALLS'].values.item()
f=sub[(sub['YEAR']==year)&(sub['AGE_GROUP']==age)&(sub['GENDER']=='F')]['NB_CALLS'].values.item()
df2=df2.append({'YEAR':year,'AGE_GROUP':age,'MALE':m,'FEMALE':f},ignore_index=True)
save(df2,'pop_pyramid.csv')
def sites_per_kaza(df, o_path):
o = pd.DataFrame()
kazas = np.unique(df['KAZA'])
for year in years:
for kaza in kazas:
sub = df[(df['KAZA'] == kaza) & (df['YEAR'] == year)]
o = o.append({
'YEAR': year,
'KAZA': kaza,
'NB_SITES': len(sub)
},
ignore_index=True)
save(o, o)
def mb(df, o_path):
df = df.groupby(['MOUHAFAZA', 'YEAR', 'KAZA'], as_index=False).agg({
'UPLINK_MB':
'sum',
'DOWNLINK_MB':
'sum',
'TOTAL_MB':
'sum',
})
df['ALFA_UP/DOWN'] = df['ALFA_UPLINK_MB'].astype(
float) / df['ALFA_DOWNLINK_MB'].astype(float)
df['ALFA_UP/TOTAL'] = df['ALFA_UPLINK_MB'].astype(
float) / df['ALFA_TOTAL_MB'].astype(float)
df['ALFA_DOWN/TOTAL'] = df['ALFA_DOWNLINK_MB'].astype(
float) / df['ALFA_TOTAL_MB'].astype(float)
save(df, o_path)
def indicators():
df = read_output('daily.csv')
df2 = pd.DataFrame()
for year in np.unique(df['YEAR']):
sub = df[df['YEAR'] == year]
ramadan = sub[sub['RAMADAN']]['NB_CALLS'].mean() / sub[
sub['RAMADAN'] == False]['NB_CALLS'].mean()
weekend = sub[sub['TYPE'] == 'WEEKEND']['NB_CALLS'].mean() / sub[
sub['TYPE'] == 'WEEKDAY']['NB_CALLS'].mean()
holiday = sub[sub['HOLIDAY']]['NB_CALLS'].mean() / sub[
sub['HOLIDAY'] == False]['NB_CALLS'].mean()
df2 = df2.append(
{
'YEAR': year,
'RAMADAN_RATIO': ramadan,
'HOLIDAY_RATIO': holiday,
'WEEKEND_RATIO': weekend
},
ignore_index=True)
save(df2, 'refugee_activity.csv')
def economics(in_df, out_df, o_path):
in_to_out = pd.DataFrame()
for _, row in in_df.iterrows():
dic = {}
dic['YEAR'] = row['YEAR']
dic['MOUHAFAZA'] = row['MOUHAFAZA']
dic['KAZA'] = row['KAZA']
out = out_df[(out_df['KAZA'] == row['KAZA'])
& (out_df['YEAR'] == row['YEAR'])]
dic['NB_RATIO'] = out['NB_CALLS'].values.item() / row['NB_CALLS']
dic['DURATION_RATIO'] = out['DURATION_IN_SEC'].values.item(
) / row['DURATION_IN_SEC']
dic['AVG_IN_DURATION_SEC'] = row['DURATION_IN_SEC'] / row['NB_CALLS']
dic['AVG_IN_DURATION_MIN'] = row['DURATION_IN_MIN'] / row['NB_CALLS']
dic['AVG_OUT_DURATION_SEC'] = out['DURATION_IN_SEC'].values.item(
) / out['NB_CALLS'].values.item()
dic['AVG_OUT_DURATION_MIN'] = out['DURATION_IN_MIN'].values.item(
) / out['NB_CALLS'].values.item()
in_to_out = in_to_out.append(dic, ignore_index=True)
save(in_to_out, o_path)
def set_weekday():
df = read('daily.csv')
df = df.apply(lambda row: weekday(row), axis=1)
save(df, 'daily.csv')
def annual_mobility():
mob = read_output('population.csv')
mob = mob.apply(lambda row: get_variation(mob, row), axis=1)
save(mob, 'mobility.csv')