def plot_distance_trip_time(df):
num_rows = df.shape[0]
title = 'trip duration v distance travelled'
print ggplot(df, aes(s.TRIP_DURATION_COL, s.DISTANCE_TRAVELED_COL_NAME)) + \
ggtitle(_make_title(title, num_rows))+ \
stat_smooth(colour="red") + \
geom_point(colour='steelblue') + \
scale_x_continuous(
# breaks=[10,20,30],
#labels=["horrible", "ok", "awesome"]
)
return df
def plot(self, inputs):
"""Plot the given X and Y axes on a scatter plot"""
if inputs.year not in self.dat.Year.values:
return
if inputs.xvar not in self.dat or inputs.yvar not in self.dat:
return
subdat = self.dat[self.dat.Year == inputs.year]
p = ggplot(subdat, aes(x=inputs.xvar, y=inputs.yvar))
p = p + geom_point()
if inputs.shownames:
p = p + geom_text(aes(label=self.ID_col), vjust=1, hjust=1)
if inputs.linear:
p = p + stat_smooth(color="red", method="lm")
return p
if (reward == 1):
wins_for_player_1[i] += 1.0
elif (reward == 0.5):
draw_for_players[i] += 1.0
print(i, wins_for_player_1[i], draw_for_players[i])
data.append({
'Type': 0,
'Wins': wins_for_player_1[i],
'Training': training_steps * (i - 1)
})
data.append({
'Type': 1,
'Wins': draw_for_players[i],
'Training': training_steps * (i - 1)
})
learnitMC(training_steps, epsilon, alpha, n)
# learnit(training_steps, epsilon, alpha) # the original learning code.
# Pandas gives you the power of R
learningdf = pd.DataFrame(data)
# I use ggplot when I generate figures in R and would like to use it with Python, HOWEVER:
# latest Pandas causes problems for ggplot so I needed these two patches:
# https://stackoverflow.com/questions/50591982/importerror-cannot-import-name-timestamp/52378663
# https://github.com/yhat/ggpy/issues/612
p = gg.ggplot(gg.aes(x='Training', y='Wins', group='Type'), data=learningdf)+ gg.xlab('Learning games') + \
gg.ylab('Wins for player 1') + gg.ggtitle("n="+str(n)) + gg.geom_point() + gg.stat_smooth(method='loess')
p.make()
filename = "experiment_" + str(n) + ".pdf"
p.save(filename)
slope = 0.3
x = randn(num) * 50. + 150.0
y = randn(num) * 5 + x * slope
plt.scatter(x, y, c='b')
# In[72]:
# plt.scatter(x[(y < 1) & (y > -1)], y[(y < 1) & (y > -1)], c='r')
# np.argsort, np.sort, complicated index slicing
dframe = pd.DataFrame({'x': x, 'y': y})
g = sns.jointplot('x', 'y', data=dframe, kind="reg")
# ## Grab Python version of ggplot http://ggplot.yhathq.com/
# In[73]:
from ggplot import ggplot, aes, geom_line, stat_smooth, geom_dotplot, geom_point
# In[74]:
ggplot(aes(x='x', y='y'), data=dframe) + geom_point() + stat_smooth(colour='blue', span=0.2)
# In[ ]:
def plot_transmission_results(tx_results, percentage_decline, save_path,
path_names):
#%% what are inputs?
# transmission results
# There'll be a folder called 'Runs prepared for ...'
# all the folders inside that folder will have a CEPAC results folder.
# tx_data is a dictionary and will have two keys, 'monthly' and 'popstats'
# 'monthly' key will only have primary transmissions data
tx_data = deepcopy(tx_results)
t = 120
total_var = 3
total_val = 4
# percentage decline
# this is also dictionary of percentage decline values for each folder
# having cepac results
# save_path eaxact folder where you want to save your images
# path_names will have paths to transmissions and sensitivity directories
#%% plot percentage decline
# geberate an environment object first
# lets go for line plot
data_plot = pd.DataFrame(
columns=['x', 'Percentage decline', 'Transmissions', 'Variable'],
index=range(0, total_var * total_val))
data_in = pd.read_excel(
os.path.join(path_names['transmission'], 'Input files',
'transmission_rate_multiplier_required_inputs.xlsx'))
col = [
'Incidence rate per 100 PY specific to high-risk group 1',
'HIV uninfected individuals in high-risk group 1',
'HIV infected individuals in high-risk group 1'
]
col_adj = ['Incidence', 'Uninfected', 'Infected']
data_in[col[0]] = data_in[col[0]].round(1)
base_val = [np.float64(0.9), 2960000, 136400]
y1_values = {col[0]: [], col[1]: [], col[2]: []}
for var in percentage_decline:
if 'HIV+' in var:
y1_values[col[2]].append(percentage_decline[var])
elif 'HIV-' in var:
y1_values[col[1]].append(percentage_decline[var])
elif 'Incidence' in var:
y1_values[col[0]].append(percentage_decline[var])
for i in range(len(col)):
idx = data_in.loc[data_in.loc[:, col[i]] != base_val[i],
col[i]].index.values[0]
data_plot.loc[idx - 1:idx + 3 - 1, 'x'] = data_in.loc[idx:idx + 3,
col[i]].values
data_plot.loc[idx - 1:idx + 3 - 1, 'Variable'] = col_adj[i]
data_plot.loc[idx - 1:idx + 3 - 1,
'Percentage decline'] = y1_values[col[i]]
# plot
df_float = data_plot.loc[data_plot.loc[:, 'Percentage decline'] <= 200, :]
(ggplot(aes(x='x', y='Percentage decline'), df_float) + geom_line() +
facet_wrap('Variable', scales='free')).save(
os.path.join(save_path, 'Percentage decline'))
del df_float
#%% visualizing transmissions
# index = range(time * number of values for each variable * number of variables)
def set_abc(run, var_idx, var_name, var_value_idx):
# set variable names
data_plot_tx.loc[(var_idx - 1) * t:((var_idx - 1) * t) + t - 1,
'Variable'] = var_name
# set variable value
data_plot_tx.loc[(var_idx - 1) * t:((var_idx - 1) * t) + t - 1,
'Value'] = data_plot.loc[
data_plot.loc[:, 'Variable'] == var_name,
'x'].values[var_value_idx]
if 'RunA' in run:
data_plot_tx.loc[(var_idx - 1) * t:((var_idx - 1) * t) + t - 1,
'RunA tx'] = tx_data[var]['monthly'][run][
'transmissions'].iloc[0:t].values
elif 'RunB' in run:
data_plot_tx.loc[(var_idx - 1) * t:((var_idx - 1) * t) + t - 1,
'RunB tx'] = tx_data[var]['monthly'][run][
'transmissions'].iloc[0:t].values
elif 'RunC' in run:
data_plot_tx.loc[(var_idx - 1) * t:((var_idx - 1) * t) + t - 1,
'RunC tx'] = tx_data[var]['monthly'][run][
'transmissions'].iloc[0:t].values
data_plot_tx = pd.DataFrame(
index=range(t * total_var * total_val),
columns=['Variable', 'Value', 'RunA tx', 'RunB tx', 'RunC tx'])
var_idx = -1
var_val_idx = [-1, -1, -1]
for var in tx_data:
var_idx += 1
if 'HIV+' in var:
var_val_idx[2] += 1
var_name = col_adj[2]
for run in tx_data[var]['monthly']:
set_abc(run, var_idx, var_name, var_val_idx[2])
elif 'HIV-' in var:
var_val_idx[1] += 1
var_name = col_adj[1]
for run in tx_data[var]['monthly']:
set_abc(run, var_idx, var_name, var_val_idx[1])
elif 'Incidence' in var:
var_val_idx[0] += 1
var_name = col_adj[0]
for run in tx_data[var]['monthly']:
set_abc(run, var_idx, var_name, var_val_idx[0])
else:
continue
data_plot_tx['t'] = 0
t_float = -1
for row in data_plot_tx.index:
if t_float == t - 1:
t_float = -1
t_float += 1
data_plot_tx.loc[row, 't'] = t_float
#%% plots for individual runs
run_col = ['RunA tx', 'RunB tx', 'RunC tx']
inci = data_plot_tx.loc[data_plot_tx.loc[:, 'Variable'] == 'Incidence', :]
inf = data_plot_tx.loc[data_plot_tx.loc[:, 'Variable'] == 'Infected', :]
uninf = data_plot_tx.loc[data_plot_tx.loc[:,
'Variable'] == 'Uninfected', :]
for i in run_col:
(ggplot(aes(x='t', y=i, color='Value'), data_plot_tx) + geom_line() +
facet_wrap('Variable', scales='free')).save(
os.path.join(
save_path,
str(i + r'_transmissions for all variable all values')))
(ggplot(aes(x='t', y=i), inci) + geom_line() +
facet_wrap('Variable', 'Value', scales='free')).save(
os.path.join(
save_path,
str(i + r'_plots for individual values of incidence')))
(ggplot(aes(x='t', y=i), inf) + geom_line() +
facet_wrap('Variable', 'Value', scales='free')).save(
os.path.join(
save_path,
str(i +
r'_plots for individual values of infected population')))
(ggplot(aes(x='t', y=i), uninf) + geom_line() +
facet_wrap('Variable', 'Value', scales='free')).save(
os.path.join(
save_path,
str(i +
'_plots for individual values of uninfected population')))
#%% compare runs ABC
data_plot_abc = {}
for var in col_adj:
float_df = pd.DataFrame(index=range(0, t * total_var * total_val),
columns=['t', 'Value', 'Transmissions', 'Run'])
insert_idx = -1
for val in data_plot.loc[data_plot.loc[:, 'Variable'] == var, 'x']:
var_df = data_plot_tx.loc[data_plot_tx.loc[:,
'Variable'] == var, :]
var_df = var_df.reset_index(drop=True)
var_val_df = var_df.loc[var_df.loc[:, 'Value'] == val, :]
var_val_df = var_val_df.reset_index(drop=True)
for c in ['RunA tx', 'RunB tx', 'RunC tx']:
insert_idx += 1
float_df.loc[insert_idx * t:(insert_idx * t) + t - 1,
'Run'] = c
float_df.loc[insert_idx * t:(insert_idx * t) + t - 1,
'Transmissions'] = var_val_df.loc[:, c].values
float_df.loc[insert_idx * t:(insert_idx * t) + t - 1,
'Run'] = c
float_df.loc[insert_idx * t:(insert_idx * t) + t - 1,
'Value'] = val
float_df.loc[insert_idx * t:(insert_idx * t) + t - 1,
't'] = np.arange(t)
data_plot_abc[var] = float_df.dropna()
(ggplot(aes(x='t', y='Transmissions', color='Run'), float_df) +
geom_line() + facet_wrap('Value', scales='free') + ggtitle(var)).save(
os.path.join(
save_path,
str(var + '_comparison of transmissions in runs ABC')))
#%% compare runs BC
for var in data_plot_abc:
float_df = data_plot_abc[var].loc[
data_plot_abc[var].loc[:, 'Run'] != 'RunA tx', :]
(ggplot(aes(x='t', y='Transmissions', color='Run'), float_df) +
geom_line(alpha=0.2) + facet_wrap('Value', scales='free') +
stat_smooth(method='loess', se=False) + ggtitle(var)).save(
os.path.join(save_path,
str(var +
'_comparison of transmissions in runs BC')))
return
def _plot_scat_w_line(self, gp_aes):
return gp_aes + gp.geom_point(color='coral') + gp.stat_smooth(span=.2, color='blue',
se=False) + gp.theme_seaborn(
context='talk')
for x in repeatedKnnResults],
columns = ['p',
'k',
'cvAccuracy',
'testAccuracy'])
ggdata = pandas.concat(
[DataFrame({'log10(p)' : log10(knnResultsSimplified.p),
'k' : knnResultsSimplified.k.apply(int),
'type' : 'cv',
'Accuracy' : knnResultsSimplified.cvAccuracy}),
DataFrame({'log10(p)' : log10(knnResultsSimplified.p),
'k' : knnResultsSimplified.k.apply(int),
'type' : 'test',
'Accuracy' : knnResultsSimplified.testAccuracy})],
axis = 0
)
ggobj = ggplot.ggplot(
data = ggdata,
aesthetics = ggplot.aes(x='log10(p)', y='Accuracy',
color='type', group='type', linetype='type')
)
ggobj += ggplot.theme_bw()
# ggobj += ggplot.scale_x_log()
ggobj += ggplot.geom_point(alpha=0.6)
ggobj += ggplot.stat_smooth()
ggobj += ggplot.facet_wrap('k')
print(ggobj)
import ggplot
from ggplot import aes, meat, geom_line, stat_smooth
ggplot(aes(x='date', y='beef'), data=meat) +\
geom_line() +\
stat_smooth(colour='blue', span=0.2)
''' ggplot(diamonds, aes(x='carat', y='price', color='cut')) +\
geom_point() +\
scale_color_brewer(type='diverging', palette=4) +\
xlab("Carats") + ylab("Price") + ggtitle("Diamonds")
ggplot(diamonds, aes(x='price', fill='cut')) +\
geom_density(alpha=0.25) +\
facet_wrap("clarity") '''
import ggplot as gp
import pandas as pd
import numpy as np
crime = pd.read_csv('crimeRatesByState2005.csv')
# 去除全美平均值和华盛顿特区两个数据点
crime2 = crime[crime.state != 'United States']
crime2 = crime2[crime.state != 'District of Columbia']
print(
gp.ggplot(gp.aes(x='murder', y='burglary'), data=crime2) +
gp.geom_point() + gp.stat_smooth(method='loess', color='red'))
