13hannes11/UU_NCML_Project
1
0
Fork 0
mirror of https://github.com/13hannes11/UU_NCML_Project.git synced 2024-09-03 20:50:59 +02:00
Code Issues Projects Releases Wiki Activity

polish plotting for voting analysis

Browse Source
This commit is contained in:
Hannes Kuchelmeister
2021-05-20 13:57:22 +02:00
parent 7fd22ccbc6
commit f3fecfdc12
1 changed files with 10 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
voting_lib/voting_analysis.py
View File

@@ -27,6 +27,8 @@ def train_model(X, grid_h, grid_w, radius, step, ep):
return sofmnet return sofmnet
def predict(model, data, grid_h, grid_w, comparison_data=pd.DataFrame()): def predict(model, data, grid_h, grid_w, comparison_data=pd.DataFrame()):
# default tight layout
plt.rcParams["figure.autolayout"] = True
X = data[:,2:] X = data[:,2:]
@@ -36,6 +38,7 @@ def predict(model, data, grid_h, grid_w, comparison_data=pd.DataFrame()):
# Plot hit map # Plot hit map
plot_hits(prediction, grid_w, grid_h) plot_hits(prediction, grid_w, grid_h)
plt.title("Hitmap")
# converting to x and y coordinates # converting to x and y coordinates
ys, xs = np.unravel_index(np.argmax(prediction, axis=1), (grid_h, grid_w)) ys, xs = np.unravel_index(np.argmax(prediction, axis=1), (grid_h, grid_w))
@@ -43,6 +46,7 @@ def predict(model, data, grid_h, grid_w, comparison_data=pd.DataFrame()):
# plotting mps # plotting mps
party_affiliation = data[:,1] party_affiliation = data[:,1]
plot_mps(data[:,0], xs, ys, party_affiliation, randomize_positions=True) plot_mps(data[:,0], xs, ys, party_affiliation, randomize_positions=True)
plt.title("Members of Parliament")
plt.show() plt.show()
# calculating party positions based on mps # calculating party positions based on mps
@@ -64,26 +68,27 @@ def predict(model, data, grid_h, grid_w, comparison_data=pd.DataFrame()):
plt.yticks(np.arange(0, grid_h+1, 1.0)) plt.yticks(np.arange(0, grid_h+1, 1.0))
plt.grid(True) plt.grid(True)
plt.title(f'Learning Radius: {model.learning_radius}, Grid Size: {grid_w}') plt.title(f'Learning Radius: {model.learning_radius}, Grid Size: {grid_w}')
plt.show()
# plotting party distances in output space # plotting party distances in output space
part_distance_out = calc_party_distances(party_pos) part_distance_out = calc_party_distances(party_pos)
plot_party_distances(part_distance_out) plot_party_distances(part_distance_out)
plt.title('Party Distances')
plt.show() plt.show()
if not comparison_data.empty: if not comparison_data.empty:
plot_parties(comparison_data, randomize_positions=False, new_plot=True) plot_parties(comparison_data, randomize_positions=False, new_plot=True)
plt.title("political compass") plt.title("Political Compass")
plt.ylabel("libertarian - authoritarian") plt.ylabel("libertarian - authoritarian")
plt.xlabel("left < economic > right") plt.xlabel("left < economic > right")
plt.show()
comparison_data_dist = calc_party_distances(comparison_data) comparison_data_dist = calc_party_distances(comparison_data)
plot_party_distances(comparison_data_dist) plot_party_distances(comparison_data_dist)
plt.show() plt.title("Political Compass Party Distances")
err = remove_NaN_rows_columns(normalize_df(part_distance_out) - normalize_df(comparison_data_dist)) err = remove_NaN_rows_columns(normalize_df(part_distance_out) - normalize_df(comparison_data_dist))
err = err * err err = err * err
plot_party_distances(err) plot_party_distances(err)
plt.title(f'distance squared error, with mse={np.nanmean(err.to_numpy()):.2f}') plt.title(f'Normalized Distance Squared Error, with MSE={np.nanmean(err.to_numpy()):.2f}')
plt.show() plt.show()