for i in range(mat.shape[1]):
for j in range(mat.shape[0]):
val = mat[j, i]
attribute = attribute_names[i]
candidate = candidate_names[j]
if val >= tresh:
pos[candidate].append(attribute)
elif val <= (-tresh):
neg[candidate].append(attribute)
return pos, neg
session = start_session("sqlite:///database.sqlite")
data_counties = load_counties(session, filter_empty=True)
counties_mat = np.array(
list(map(lambda x: x.get_vector().numerical(), data_counties)))
vector_labels = data_counties[0].get_vector().labels(short=True)
m = np.corrcoef(counties_mat, rowvar=False)
candidates_mat = m[:7, 7:]
dem_mat = candidates_mat[:2, :]
rep_mat = candidates_mat[2:, :]
democrat_candidate_names = vector_labels[:2]
republican_candidate_names = vector_labels[2:7]
attribute_names = vector_labels[7:]
from dbload import load_counties, load_states, start_session
from primaryarea import CountyArea
import numpy as np
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
session = start_session("sqlite:///database.sqlite")
data_counties = load_counties(session)
counties_mat = np.array(
list(map(lambda x: x.get_vector().numerical(), data_counties)))
data_states = load_states(session)
states_mat = np.array(
list(map(lambda x: x.get_vector().numerical(), data_states)))
vector_labels = data_counties[0].get_vector().labels(short=True)
candidate_names = vector_labels[:7]
counties_names = list(map(lambda x: x.facts.area_name, data_counties))
states_names = list(map(lambda x: x.facts.area_name, data_states))
election_results = counties_mat[:, :7]
attributes = counties_mat[:, 7:]
reg = LinearRegression(positive=True).fit(attributes, election_results)
actual_state_results = states_mat[:, :7]
state_attributes = states_mat[:, 7:]
predicted_state_results = reg.predict(state_attributes)
for i, candidate_name in enumerate(candidate_names):
actual_results = actual_state_results[:, i]