def test_deciles_chart():
"""Currently just tests it doesn't fail
"""
df = pd.DataFrame(np.random.rand(1000, 1), columns=["val"])
months = pd.date_range("2018-01-01", periods=12, freq="M")
df["month"] = np.random.choice(months, len(df))
charts.deciles_chart(
df,
period_column="month",
column="val",
title="thing",
show_outer_percentiles=True,
)
def plot_charts(out, outer_percentiles):
charts.deciles_chart(
out,
period_column='month',
column='value',
#title=Title,
ylabel="rate per 1000",
show_outer_percentiles=outer_percentiles,
show_legend=True)
# Now add a single line against the deciles
#df_subject = pd.DataFrame(np.random.normal(0.6, 0.05, 37), columns=['val'])
#df_subject['month'] = out["month"].drop_duplicates()
#df_subject.set_index('month')
#plt.plot(df_subject['month'], df_subject['val'], 'r--')
plt.show()
right_plot = plt.subplot(layout[1], sharey=ax1) # Share the Y axis
# ...and because it's shared, suppress ticks on the second chart
plt.setp(right_plot.get_yticklabels(), visible=False)
# make a datafrom with a date column and values columns
df = pd.DataFrame(np.random.rand(1000, 2), columns=["val1", "val2"])
months = pd.date_range("2018-01-01", periods=12, freq="M")
df["month"] = np.random.choice(months, len(df))
# Plot a deciles chart on the left axis
charts.deciles_chart(
df,
period_column="month",
column="val1",
title="Random values",
ylabel="n",
show_outer_percentiles=True,
show_legend=False,
ax=left_plot,
)
# ...and one on the right axis
charts.deciles_chart(
df,
period_column="month",
column="val2",
title="Random values 2",
show_outer_percentiles=True,
show_legend=True,
ax=right_plot,
)
ax = df_zop_breach.groupby(["month"])['items_breach'].sum().plot(
kind='line',
title=
"Total number of items for zopiclone that had greater than 31 days qty")
ax.axvline(pd.to_datetime('2014-06-01'), color='black', linestyle='--',
lw=2) ##law change
plt.ylim(0, )
# +
#create sample deciles
charts.deciles_chart(
df_zop_breach,
period_column='month',
column='percent_qty_breach',
title=
"Proportion of one month quantity breaches of zopiclone (Devon - CCG) ",
show_outer_percentiles=True)
#add in example CCG (Devon - 15N)
df_subject = df_zop_breach.loc[df_zop_breach['pct'] == '15N'].sort_values(
by=['month'])
plt.plot(df_subject['month'], df_subject['percent_qty_breach'], 'r--')
plt.show()
# -
##Restrict to latest month (may-2019) to create a map
may_df_zop = df_zop_breach.loc[(df_zop_breach['month'] == '2020-05-01')]
df_amp_recommended_ccg,
how='left',
left_on=['month', 'pct'],
right_on=['month', 'pct'],
suffixes=("_generic_rx", "_all_rx_rec_amp"))
df_measure["measure_value"] = 100 * (
df_measure.total_items_generic_rx /
df_measure.total_items_all_rx_rec_amp).fillna(0)
df_measure.head()
# +
charts.deciles_chart(
df_measure,
period_column='month',
column='measure_value',
title=
" CCG variation - Generic prescriptions \n in breach of dm+d brand prescribing recommendations",
ylabel=" % ",
show_outer_percentiles=True,
show_legend=True)
#add in example CCG (Devon)
df_subject = df_measure.loc[df_measure['pct'] == '15N']
plt.plot(df_subject['month'], df_subject['measure_value'], 'r--')
plt.ylim(0, 100)
plt.show()
# -
#create choropeth map
plt.figure(figsize=(12, 7))
df_list = pd.read_csv('listsize.csv')
df_list['month'] = df_list['month'].astype('datetime64[ns]')
df_list.head(5)
top_abx_1000 = pd.merge(ccg_total_abx, df_list, on=['month', 'pct'])
top_abx_1000['top_abx_items_per_1000'] = 1000 * (top_abx_1000['items'] /
top_abx_1000['list_size'])
top_abx_1000.head(5)
# +
#create sample deciles & prototype measure
charts.deciles_chart(
top_abx_1000,
period_column='month',
column='top_abx_items_per_1000',
title=
"Topical antibiotics items per 1000 (Bath and North East Somerset CCG) ",
show_outer_percentiles=True)
#add in example CCG (Islington)
df_subject = top_abx_1000.loc[top_abx_1000['pct'] == '11E']
plt.plot(df_subject['month'], df_subject['top_abx_items_per_1000'], 'r--')
plt.show()
# -
#create choropeth map of items per 1000 patients
plt.figure(figsize=(12, 7))
latest_top_abx_1000 = top_abx_1000.loc[(top_abx_1000['month'] >= '2019-01-01')
&
ccg_total_abx = df_inj_abx.groupby(["month",
"pct"])["items"].sum().reset_index()
ccg_total_abx.head(5)
df_abx_1000 = pd.merge(ccg_total_abx, df_list, on=['month', 'pct'])
df_abx_1000['inj_abx_items_per_1000'] = 1000 * (df_abx_1000['items'] /
df_abx_1000['list_size'])
df_abx_1000.head(5)
# +
#create sample deciles & prototype measure
charts.deciles_chart(
df_abx_1000,
period_column='month',
column='inj_abx_items_per_1000',
title=
"Injectable antibiotics items per 1000 (Bath and North East Somerset CCG) ",
show_outer_percentiles=False)
#add in example CCG (Islington)
df_subject = df_abx_1000.loc[df_abx_1000['pct'] == '11E']
plt.plot(df_subject['month'], df_subject['inj_abx_items_per_1000'], 'r--')
plt.show()
# -
#create choropeth map of cost per 1000 patients
plt.figure(figsize=(12, 7))
latest_df_abx_1000 = df_abx_1000.loc[(df_abx_1000['month'] >= '2019-01-01')
& (df_abx_1000['month'] <= '2019-12-01')]
#create only latest month's CCG data
latest_date = df_ccg['month'].max()
df_ccg_map = df_ccg.loc[df_ccg['month'] == latest_date]
ld_string = latest_date.month_name() + " " + str(
latest_date.year) #this creates string to label chart
plt.figure(figsize=(20, 7))
maps.ccg_map(df_ccg_map,
title='Percentage of DOACs and warfarin prescribed as DOACS\n(' +
ld_string + ')',
column='perc_doacs',
separate_london=True,
plot_options={
'vmax': 100,
'vmin': 0,
'cmap': 'coolwarm'
})
plt.show()
# ## Deciles of CCGs for percentage as DOACs
#plot deciles for DOACs
charts.deciles_chart(
df_ccg,
period_column='month',
column='perc_doacs',
title="CCGs - Percentage of DOACs and warfarin \nprescribed as DOACS",
show_outer_percentiles=True)
plt.show()
df_list = bq.cached_read(sql2, csv_path=os.path.join('..','data','list_size.csv'))
df_list['month'] = df_list['month'].astype('datetime64[ns]')
df_list.head(3)
ccg_total = df_poisoning.groupby(["month", "pct"])["total_items"].sum().reset_index()
ccg_total.head()
poisoning_ccg_1000 = pd.merge(ccg_total, df_list, on=['month', 'pct'])
poisoning_ccg_1000['items_per_1000'] = 1000* (poisoning_ccg_1000['total_items']/poisoning_ccg_1000['list_size'])
poisoning_ccg_1000.head(3)
# +
#create sample deciles & prototype measure
charts.deciles_chart(
poisoning_ccg_1000,
period_column='month',
column='items_per_1000',
title="Poisoning items per 1000 (Islington CCG) ",
show_outer_percentiles=False)
#add in example CCG (Islington)
df_subject = poisoning_ccg_1000.loc[poisoning_ccg_1000['pct'] == '08H']
plt.plot(df_subject['month'], df_subject['items_per_1000'], 'r--')
plt.show()
# +
#create choropeth map of cost per 1000 patients
plt.figure(figsize=(12, 7))
latest_poisoning_df_1000 = poisoning_ccg_1000.loc[(poisoning_ccg_1000['month'] >= '2019-04-01') & (poisoning_ccg_1000['month'] <= '2020-02-01')]
plt = maps.ccg_map(latest_poisoning_df_1000, title="Poisoning items per 1000 \n Apr 2019 - Feb 2020 ", column='items_per_1000', separate_london=True)
# +
from ebmdatalab import charts
import matplotlib.gridspec as gridspec
import importlib
importlib.reload(charts)
# make a datafrom with a date column and a values column
df = pd.DataFrame(np.random.rand(1000, 1), columns=["val"])
months = pd.date_range("2018-01-01", periods=12, freq="M")
df["month"] = pd.to_datetime(np.random.choice(months, len(df)))
charts.deciles_chart(
df,
period_column="month",
column="val",
title="Random values",
ylabel="n",
show_outer_percentiles=True,
show_legend=True,
)
# Now add a single line against the deciles
df_subject = pd.DataFrame(np.random.rand(12, 1), columns=["val"]) * 100
df_subject["month"] = months
df_subject.set_index("month")
plt.plot(df_subject["month"], df_subject["val"], "r--")
plt.show()
