def initialise_g_score(self, data):
'''Initialises all parameters'''
if data == 0:
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
df = pd.DataFrame()
df['nic'] = self.year_list_decile[name_list[i]]['nic']
df['sa_finance1_cocode'] = self.year_list_decile[
name_list[i]]['sa_finance1_cocode']
df['sa_company_name'] = self.year_list_decile[
name_list[i]]['sa_company_name']
self.g_score_data_20[name_list[i]] = df
name_list = build_name_list(2007, 10)
for i in range(len(name_list)):
df = pd.DataFrame()
df['sa_finance1_cocode'] = self.year_list[
name_list[i]]['sa_finance1_cocode']
df['sa_company_name'] = self.year_list[
name_list[i]]['sa_company_name']
self.g_score_data[name_list[i]] = df
elif data == 1:
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
link_des = 'g_score_data_20/' + name_list[
i] + '_g_score_data_20.csv'
g_score_data_20 = pd.read_csv(link_des, header=None)
self.g_score_data_20[name_list[i]] = g_score_data_20
name_list = build_name_list(2007, 10)
for i in range(len(name_list)):
link_des = 'g_score_data/' + name_list[i] + '_g_score_data.csv'
g_score_data = pd.read_csv(link_des, header=None)
self.g_score_data[name_list[i]] = g_score_data
def capex(self):
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
capex_yearly = []
year_list_decile = self.year_list_decile[name_list[i]].fillna(0)
g_score_data_20 = self.g_score_data_20[name_list[i]]
for company in g_score_data_20.iloc[:, 2]:
df1 = year_list_decile[
year_list_decile['sa_company_name'].str.match(company)]
df2 = g_score_data_20[g_score_data_20[2].str.match(company)][3]
try:
capex = float(
df1['sa_building_net_addn_in_yr'].values[0]
) + float(
df1['sa_comm_equip_net_addn_in_yr'].values[0]
) + float(
df1['sa_computer_it_net_addn_in_yr'].values[0]
) + float(
df1['sa_elec_install_fitting_net_addn_in_yr'].values[0]
) + float(df1['sa_gfa_net_addn_in_yr'].values[0]) + float(
df1['sa_net_furn_social_oth_fxd_ast'].values[0]
) + float(
df1['sa_plant_net_addn_in_yr'].values[0]
) + float(df1['sa_sw_net_addn_in_yr'].values[0]) + float(
df1['sa_transport_infra_net_addn_in_yr'].values[0]
) + float(df1['sa_transport_veh_net_addn_in_yr'].values[0])
capex = capex / float(df2.values[0])
capex_yearly.append(capex)
except:
capex_yearly.append(float('NaN'))
capex_yearly = pd.DataFrame(capex_yearly)
self.g_score_data_20[name_list[i]] = pd.concat(
[self.g_score_data_20[name_list[i]], capex_yearly],
axis=1,
ignore_index=True)
def sales_var(self):
name_list = build_name_list(2007, 10)
for i in range(len(name_list) - 4):
sales_var_yearly = []
for company in self.g_score_data_20[name_list[i + 4]].iloc[:, 2]:
sales_past = []
for j in range(3):
sales_1 = self.year_list[name_list[i + j]]
df1 = sales_1[sales_1['sa_company_name'].str.match(
company)]['sa_sales'].fillna(0)
sales_2 = self.year_list[name_list[i + j + 1]]
df2 = sales_2[sales_2['sa_company_name'].str.match(
company)]['sa_sales'].fillna(0)
try:
a = (df2.values[0] - df1.values[0])
sales_past.append(a)
except:
pass
try:
sales_past = pd.DataFrame(sales_past)
if sales_past.shape[0] == 0 or sales_past.shape[0] == 1:
sales_var = float(float('inf'))
else:
sales_var = (sales_past.var()).values[0]
sales_var_yearly.append(sales_var)
except:
sales_var_yearly.append(float('inf'))
sales_var_yearly = (pd.DataFrame(sales_var_yearly))
self.g_score_data_20[name_list[i + 4]] = pd.concat(
[self.g_score_data_20[name_list[i + 4]], sales_var_yearly],
axis=1,
ignore_index=True)
def g_scores_to_csv(self):
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
self.g_scores[name_list[i]].to_csv('g_scores/' + name_list[i] +
'_g_scores.csv',
header=False,
index=False)
def roa_var(self):
name_list = build_name_list(2008, 9)
for i in range(len(name_list) - 3):
roa_var_yearly = []
for company in self.year_list_decile[name_list[
i + 3]]['sa_company_name']:
roa_past = []
for j in range(3):
g_score_data = self.g_score_data[name_list[i + j]]
df = g_score_data[g_score_data[1].str.match(company)][3]
try:
roa_past.append(df.values[0])
except:
pass
try:
roa_past = pd.DataFrame(roa_past)
if roa_past.shape[0] == 0 or roa_past.shape[0] == 1:
roa_var = float('inf')
else:
roa_var = (roa_past.var()).values[0]
roa_var_yearly.append(roa_var)
except:
roa_var_yearly.append(float('inf'))
roa_var_yearly = pd.DataFrame(roa_var_yearly)
self.g_score_data_20[name_list[i + 3]] = pd.concat(
[self.g_score_data_20[name_list[i + 3]], roa_var_yearly],
axis=1,
ignore_index=True)
def g_score_data_to_csv(self):
name_list = build_name_list(2007, 10)
for i in range(len(name_list)):
self.g_score_data[name_list[i]].to_csv(
'g_score_data/' + name_list[i] + '_g_score_data.csv',
header=False,
index=False)
def average_assets_all(self):
'''Get average assets'''
name_list = build_name_list(2007, 10)
for i in range(len(name_list) - 1):
average_assets_yearly = []
for company in self.year_list[name_list[i + 1]]['sa_company_name']:
df = self.year_list[name_list[i + 1]]
assets_t1 = df[df['sa_company_name'].str.match(
company)]['sa_total_assets']
df = self.year_list[name_list[i]]
assets_t0 = df[df['sa_company_name'].str.match(
company)]['sa_total_assets']
try:
average_assets = (float(assets_t0.values[0]))
average_assets_yearly.append(average_assets)
except:
try:
average_assets = (float(assets_t1.values[0]))
average_assets_yearly.append(average_assets)
except:
average_assets_yearly.append(float('NaN'))
average_assets_yearly = pd.DataFrame(average_assets_yearly)
self.g_score_data[name_list[i + 1]] = pd.concat(
[self.g_score_data[name_list[i + 1]], average_assets_yearly],
axis=1,
ignore_index=True)
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
average_assets_yearly_20 = []
#link_des = 'g_score_data/'+name_list[i]+'g_score_data.csv'
#g_score_data = pd.read_csv(link_des, header=None)
g_score_data = self.g_score_data[name_list[i]]
for company_20 in self.year_list_decile[
name_list[i]]['sa_finance1_cocode']:
crap = []
crap.append(str(company_20))
df = g_score_data[g_score_data[0].isin(crap)][2]
average_assets_20 = float(df.values[0])
average_assets_yearly_20.append(average_assets_20)
average_assets_yearly_20 = pd.DataFrame(average_assets_yearly_20)
self.g_score_data_20[name_list[i]] = pd.concat(
[self.g_score_data_20[name_list[i]], average_assets_yearly_20],
axis=1,
ignore_index=True)
def __init__(self):
'''Returns pandas file of all year shit'''
year_list_decile = {}
year_list = {}
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
year_list_decile[
name_list[i]] = pd.read_csv('cleaned_data/clean_sorted/' +
name_list[i] + '_clean_sorted.csv')
name_list = build_name_list(2007, 10)
for i in range(len(name_list)):
year_list[name_list[i]] = pd.read_csv('data_collected/data/' +
name_list[i] + '_yearly.csv')
self.year_list_decile = year_list_decile
self.year_list = year_list
self.g_score_data_20 = {}
self.g_score_data = {}
self.g_scores = {}
def roa(self):
name_list = build_name_list(2008, 9)
for i in range(len(name_list)):
roa_yearly = []
year_list = self.year_list[name_list[i]]
g_score_data = self.g_score_data[name_list[i]]
for company in self.year_list[name_list[i]]['sa_company_name']:
df1 = year_list[year_list['sa_company_name'].str.match(
company)].fillna(0)
df2 = g_score_data[g_score_data[1].str.match(company)][2]
try:
roa = (float(df1['sa_pat'].values[0]) +
float(df1['sa_extra_ordi_inc'].values[0])) / float(
df2.values[0])
roa_yearly.append(roa)
except:
roa_yearly.append(float('NaN'))
roa_yearly = pd.DataFrame(roa_yearly)
self.g_score_data[name_list[i]] = pd.concat(
[self.g_score_data[name_list[i]], roa_yearly],
axis=1,
ignore_index=True)
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
roa_yearly_20 = []
#link_des = 'g_score_data/'+name_list[i]+'g_score_data.csv'
#g_score_data = pd.read_csv(link_des, header=None)
g_score_data = self.g_score_data[name_list[i]]
for company_20 in self.year_list_decile[
name_list[i]]['sa_finance1_cocode']:
crap = []
crap.append(str(company_20))
df = g_score_data[g_score_data[0].isin(crap)][3]
roa_20 = float(df.values[0])
roa_yearly_20.append(roa_20)
roa_yearly_20 = pd.DataFrame(roa_yearly_20)
self.g_score_data_20[name_list[i]] = pd.concat(
[self.g_score_data_20[name_list[i]], roa_yearly_20],
axis=1,
ignore_index=True)
def accruals(self):
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
accruals_yearly = []
g_score_data_20 = self.g_score_data_20[name_list[i]]
for company in g_score_data_20.iloc[:, 2]:
df = g_score_data_20[g_score_data_20[2].str.match(company)]
try:
cfo = df[5].values[0]
roa = df[4].values[0]
accruals = cfo - roa
accruals_yearly.append(accruals)
except:
accruals_yearly.append(float('NaN'))
accruals_yearly = pd.DataFrame(accruals_yearly)
self.g_score_data_20[name_list[i]] = pd.concat(
[self.g_score_data_20[name_list[i]], accruals_yearly],
axis=1,
ignore_index=True)
def rnd(self):
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
rnd_yearly = []
year_list_decile = self.year_list_decile[name_list[i]]
g_score_data_20 = self.g_score_data_20[name_list[i]]
for company in g_score_data_20.iloc[:, 2]:
df1 = year_list_decile[year_list_decile['sa_company_name'].str.
match(company)]['sa_rnd_exp'].fillna(0)
df2 = g_score_data_20[g_score_data_20[2].str.match(company)][3]
try:
rnd = float(df1.values[0]) / float(df2.values[0])
rnd_yearly.append(rnd)
except:
rnd_yearly.append(float('NaN'))
rnd_yearly = pd.DataFrame(rnd_yearly)
self.g_score_data_20[name_list[i]] = pd.concat(
[self.g_score_data_20[name_list[i]], rnd_yearly],
axis=1,
ignore_index=True)
def cfo(self):
name_list = build_name_list(2011, 6)
for i in range(len(name_list)):
cfo_yearly = []
year_list_decile = self.year_list_decile[name_list[i]]
g_score_data_20 = self.g_score_data_20[name_list[i]]
for company in g_score_data_20.iloc[:, 2]:
df1 = year_list_decile[
year_list_decile['sa_company_name'].str.match(
company)]['sa_cf_net_frm_op_activity']
df2 = g_score_data_20[g_score_data_20[2].str.match(company)][3]
try:
cfo = float(df1.values[0]) / float(df2.values[0])
cfo_yearly.append(cfo)
except:
cfo_yearly.append(float('NaN'))
cfo_yearly = pd.DataFrame(cfo_yearly)
self.g_score_data_20[name_list[i]] = pd.concat(
[self.g_score_data_20[name_list[i]], cfo_yearly],
axis=1,
ignore_index=True)
def g_score_calc(self):
g_scores = {}
name_list = build_name_list(2011, 6)
for z in range(len(name_list)):
g_score = self.g_score_data_20[name_list[z]].fillna(float('inf'))
for j in range(g_score.shape[1] - 4):
ref = 0
ref_index = -1
first = 0
g_score_int = g_score.reset_index(drop=True)
n = g_score.shape[0]
for index, nic in g_score.iloc[:, 0].iteritems():
if ref != int(nic) or index == n - 1:
inf_count = 0
if first == 0:
g = g_score.iloc[ref_index:index, :]
first = first + 1
g = sort(g, j + 4, 1)
n = g.shape[0]
for i in range(n):
if j == 2:
if g.iloc[i, j + 4] > 0:
g.iloc[i, j + 4] = 1
else:
g.iloc[i, j + 4] = 0
else:
if g.iloc[i, j + 4] == float('inf'):
g.iloc[i, j + 4] = 0
inf_count = inf_count + 1
else:
if i < (n - inf_count) / 2:
g.iloc[i, j + 4] = 1
else:
g.iloc[i, j + 4] = 0
g_score_int = g
elif first == 1:
g = g_score.iloc[ref_index:index, :]
g = sort(g, j + 4, 1)
n = g.shape[0]
for i in range(n):
if g.iloc[i, j + 4] == float('inf'):
g.iloc[i, j + 4] = 0
inf_count = inf_count + 1
else:
if i < (n - inf_count) / 2:
g.iloc[i, j + 4] = 1
else:
g.iloc[i, j + 4] = 0
g_score_int = pd.concat([g_score_int, g],
axis=0,
ignore_index=True)
ref = int(nic)
ref_index = index
else:
pass
g_score = g_score_int
g_1 = []
for i in range(g_score.shape[0]):
a = 0
for j in range(7):
a = a + g_score.iloc[i, j + 4]
g_1.append(a)
g_score = pd.concat([g_score, pd.DataFrame(g_1)],
axis=1,
ignore_index=True)
g_scores = sort(g_score, 12, 1).reset_index(drop=True)
g_scores_1 = g_scores.iloc[:, 0:3]
g_scores_2 = g_scores.iloc[:, 12]
self.g_scores[name_list[z]] = pd.concat([g_scores_1, g_scores_2],
axis=1,
ignore_index=True)
def get_g_scores(self, i):
name_list = build_name_list(2011, 6)
print(name_list[i])
print(self.g_scores[name_list[i]])
def get_g_score_data_20(self, i):
name_list = build_name_list(2011, 6)
print(name_list[i])
print(self.g_score_data_20[name_list[i]])
def get_g_score_data(self, i):
name_list = build_name_list(2007, 10)
print(name_list[i])
print(self.g_score_data[name_list[i]])
