def make_japan_heatmap(filename, title, npa1d, populations):
""" 都道府県別ヒートマップを表示する """
fig = go.Figure()
w = 610
h = 630
cmap = plt.get_cmap("plasma")
norm = plt.Normalize(vmin=np.min(npa1d[1:]), vmax=np.max(npa1d[1:]))
fcol = lambda x: '#' + bytes(cmap(norm(x), bytes=True)[:3]).hex()
fig.add_trace(go.Scatter(x=[0, w], y=[0, h], mode='markers', marker_opacity=0))
fig.add_trace(go.Heatmap(x=[0, 0], y=[0, 0], opacity=0,
z=[np.min(npa1d[1:]), np.max(npa1d[1:])],
zmin=np.min(npa1d[1:]), zmax=np.max(npa1d[1:]),
type='heatmap', colorscale='plasma', showscale=True))
map_cols = {}
for k, v in populations.items():
map_cols[v['ja']] = fcol(npa1d[v['code']])
pict = picture(map_cols)
axis_template = lambda x: dict(
range=[0, x], autorange=False, showgrid=False, zeroline=False,
linecolor='black', showticklabels=False, ticks='')
fig.update_layout(title=title, xaxis=axis_template(w), yaxis=axis_template(h),
showlegend=False, width=w, height=h, autosize=False,
margin={"l": 0, "r": 0, "t":40, "b": 0}
)
fig.add_layout_image(dict(
x=0, sizex=w, y=h, sizey=h, xref="x", yref="y", opacity=1,
layer="below", sizing="stretch", source=Image.fromarray(pict)))
show_and_save_plotly(fig, filename, js=False, show=True, image=True, html=False)
def jpn_map(prefectures, values, title, cmap_name="Reds", filename=None):
"""
Show colored Japan prefecture map.
@prefectures <list/pd.Series[str]>: prefecture name.
@values <int/float>: value of each prefectures
@title <str>: title of the figure
@cmap_name <str>: Please refere to
- https://matplotlib.org/3.1.0/tutorials/colors/colormaps.html
@filename <str>: filename of the figure, or None (show figure)
"""
# Prefecture code created in
# https://www.kaggle.com/lisphilar/eda-of-japan-dataset
# Primary data: http://nlftp.mlit.go.jp/ksj/gml/codelist/PrefCd.html
# cf. https://www.japanvisitor.com/japan-travel/prefectures-map
pref_code_dict = dict([('Hokkaido', 1), ('Aomori', 2), ('Iwate', 3),
('Miyagi', 4), ('Akita', 5), ('Yamagata', 6),
('Fukushima', 7), ('Ibaraki', 8), ('Tochigi', 9),
('Gunma', 10), ('Saitama', 11), ('Chiba', 12),
('Tokyo', 13), ('Kanagawa', 14), ('Niigata', 15),
('Toyama', 16), ('Ishikawa', 17), ('Fukui', 18),
('Yamanashi', 19), ('Nagano', 20), ('Gifu', 21),
('Shizuoka', 22), ('Aichi', 23), ('Mie', 24),
('Shiga', 25), ('Kyoto', 26), ('Osaka', 27),
('Hyogo', 28), ('Nara', 29), ('Wakayama', 30),
('Tottori', 31), ('Shimane', 32), ('Okayama', 33),
('Hiroshima', 34), ('Yamaguchi', 35),
('Tokushima', 36), ('Kagawa', 37), ('Ehime', 38),
('Kochi', 39), ('Fukuoka', 40), ('Saga', 41),
('Nagasaki', 42), ('Kumamoto', 43), ('Oita', 44),
('Miyazaki', 45), ('Kagoshima', 46),
('Okinawa', 47)])
# Data to dataframe
df = pd.DataFrame({"Name": prefectures, "Value": values})
df["Code"] = df["Name"].map(pref_code_dict)
df["Japanese_name"] = df["Code"].apply(lambda x: japanmap.pref_names[x])
df = df.set_index("Japanese_name")
# Color code
cmap = cm.get_cmap(cmap_name)
norm = matplotlib.colors.Normalize(vmin=df["Value"].min(),
vmax=df["Value"].max())
def fcol(x):
return "#" + bytes(cmap(norm(x), bytes=True)[:3]).hex()
# Show figure
plt.xticks(color="None")
plt.yticks(color="None")
plt.tick_params(length=0)
plt.imshow(japanmap.picture(df["Value"].apply(fcol)))
mappable = cm.ScalarMappable(norm=norm, cmap=cmap)
mappable._A = []
plt.colorbar(mappable)
plt.title(title)
if filename is None:
plt.show()
return None
plt.savefig(filename, bbox_inches="tight", transparent=False, dpi=300)
plt.clf()
return None
def create_figure_y(self):
cmap = plt.get_cmap('binary')
norm = plt.Normalize(vmin=9.0, vmax=19.0)
fcol = lambda x: '#' + bytes(cmap(norm(x), bytes=True)[:3]).hex()
plt.colorbar(plt.cm.ScalarMappable(norm, cmap))
plt.title('y mean')
df = self.data['Y'].apply(fcol)
# indexを振り直す
df.index = df.index + 1
plt.imshow(picture(df))
plt.show()
plt.close()
def whitejapanmap():
import urllib
import base64
import io
import matplotlib.pyplot as plt
from japanmap import picture
plt.rcParams['figure.figsize'] = 6, 6
plt.imshow(picture())
fig = plt.gcf()
buf = io.BytesIO()
fig.savefig(buf, format='png')
buf.seek(0)
string = base64.b64encode(buf.read())
uri = urllib.parse.quote(string)
return uri
def createInfMapJapan(self, df, date):
if df is False:
return
# 空のグラフオブジェクトを作成
fig = plt.figure(figsize=(self.graph_width_map, self.graph_hight_map),
facecolor=self.bgcolor_map, dpi=self.graph_dpi,
tight_layout=True)
# 都道府県名を行名に設定
df_in = df.set_index("Pref")
# カラーマップの作成
cmap = plt.get_cmap('Blues')
# 値の正規化
norm = mpl.colors.Normalize(vmin=df_in["InfNum"].min(), vmax=df_in["InfNum"].max())
# 色変換
fcol = lambda x: '#' + bytes(cmap(norm(x), bytes=True)[:3]).hex()
# カラーバーの作成
mappable = plt.cm.ScalarMappable(norm, cmap)
mappable.set_array(df_in["InfNum"])
plt.colorbar(mappable)
# タイトルの設定
title = self.title_map + date
plt.title(title)
# 軸ラベル削除
plt.tick_params(labelbottom=False, labelleft=False,
labelright=False, labeltop=False)
# 軸メモリ削除
plt.tick_params(bottom=False, left=False, right=False, top=False)
# 描画
plt.imshow(picture(df_in["InfNum"].apply(fcol)));
# 表示
# plt.show()
return fig
def create_figure(self, mcmc_sample):
r = mcmc_sample['r'].mean(axis=0)
# print(r)
df = pandas.DataFrame(r, columns=['r'])
# print(df)
cmap = plt.get_cmap('binary')
norm = plt.Normalize(vmin=9.0, vmax=19.0)
fcol = lambda x: '#' + bytes(cmap(norm(x), bytes=True)[:3]).hex()
plt.colorbar(plt.cm.ScalarMappable(norm, cmap))
plt.title('r')
# indexを振り直す
df = df['r'].apply(fcol)
df.index = df.index + 1
plt.imshow(picture(df))
plt.show()
plt.close()
equal_to(sum_poly([q[i][c] for c in range(num_colors)]), 1)
for i in range(num_region)
]
# 隣接する領域間の制約
adj_constraints = [
# 都道府県コードと配列インデックスは1ずれてるので注意
penalty(q[i][c] * q[j - 1][c]) for i in range(num_region)
for j in jm.adjacent(i + 1) # j: 隣接している都道府県コード
if i + 1 < j for c in range(num_colors)
]
constraints = sum(reg_constraints) + sum(adj_constraints)
model = BinaryQuadraticModel(constraints)
result = solver.solve(model)
if len(result) == 0:
raise RuntimeError("Any one of constraints is not satisfied.")
values = result[0].values
q_values = decode_solution(q, values, 1)
color_indices = np.where(np.array(q_values) == 1)[1]
color_map = {
jm.pref_names[i + 1]: colors[color_indices[i]]
for i in range(len(color_indices))
}
plt.rcParams["figure.figsize"] = 6, 6
plt.imshow(jm.picture(color_map))
plt.show()
# fmt='o', ecolor='gray', ms=10, alpha=0.8, marker='o', mfc='green', capsize=3)
ax = plt.axes()
ax.set_aspect('equal')
plt.setp(ax, xlabel='Y', ylabel='r')
plt.show()
plt.close()
if __name__ == '__main__':
m = Markov2Dim('data-map-temperature.txt', 'data-map-neighbor.txt',
'../model/model12-8')
m.describe()
# Figure of Japan map.
plt.rcParams['figure.figsize'] = 6, 6
plt.imshow(picture())
plt.show()
plt.close()
m.create_figure_y()
# モデリング
stan_data = m.create_data()
mcmc_sample = m.fit(stan_data)
# モデリングした結果の可視化
m.create_figure(mcmc_sample)
# 予測と実測のプロット
m.observe_diff_mean(mcmc_sample)
def get_svg(request, pk):
# plt.rcParams['figure.figsize'] = 6, 6
return plt.imshow(picture())
# Google Trends(トレンド)からデータ取得
pytrends = TrendReq(hl='ja-JP', tz=-540)
kw_list = [keyword]
pytrends.build_payload(kw_list, cat=0, timeframe='today 1-m', geo='JP', gprop='')
df = pytrends.interest_by_region(resolution='JP', inc_low_vol=True, inc_geo_code=True)
df['geoCode'] = df['geoCode'].str.replace('JP-', '').astype(int)
# データ作成
# 共通データ作成
cmap = plt.get_cmap('jet')
norm = plt.Normalize(vmin=df[keyword].min(), vmax=df[keyword].max())
# カラーバー用データを作成
mappable = cm.ScalarMappable(cmap=cmap, norm=norm)
mappable._A = []
# 日本地図用データを作成
fcol = lambda x: '#' + bytes(cmap(norm(x), bytes=True)[:3]).hex()
# 図を描画
# 図のベースを描く
plt.figure(figsize=(10, 8))
# 日本地図を描く
plt.imshow(picture(df[keyword].apply(fcol)));
# カラーバーを描く
plt.colorbar(mappable)
plt.savefig('{}.png'.format(keyword))
plt.show()
