def current():
cases, stats = filter_data(InsolvencyType.PRIVATE)
proceedings = cases[CaseType.VERFAHRENEROEFFNET]
# Filter for recent proceedings
latest_data = max(p["date"] for p in proceedings)
date_7_days_ago = latest_data - dt.timedelta(days=7)
date_14_days_ago = latest_data - dt.timedelta(days=14)
last_7_days = len([p for p in proceedings if p["date"] > date_7_days_ago])
the_7_days_before = len([
p for p in proceedings
if date_7_days_ago >= p["date"] > date_14_days_ago
])
try:
percent_change = f"{signed(round((last_7_days - the_7_days_before) / the_7_days_before * 100))}%"
except ZeroDivisionError:
percent_change = "+∞%"
df = pd.DataFrame(
data={
"Der letzten 7 Tage": {
"Insolvenzverfahren": last_7_days
},
"Die 7 Tage davor": {
"Insolvenzverfahren": the_7_days_before
},
"Veränderung": {
"Insolvenzverfahren": percent_change
},
})
return df
def current(case_type):
cases, stats = filter_data(InsolvencyType.REGULAR)
cases = cases[case_type]
# Filter for recent proceedings
latest_data = max(p["date"] for p in cases)
date_7_days_ago = latest_data - dt.timedelta(days=7)
date_14_days_ago = latest_data - dt.timedelta(days=14)
last_7_days = len([p for p in cases if p["date"] > date_7_days_ago])
the_7_days_before = len(
[p for p in cases if date_7_days_ago >= p["date"] > date_14_days_ago]
)
try:
percent_change = f"{signed(round((last_7_days - the_7_days_before) / the_7_days_before * 100))}%"
except ZeroDivisionError:
percent_change = "+∞%"
df = pd.DataFrame(
data={
"Der letzten 7 Tage": {CASE_TYPE_HEADERS[case_type]: last_7_days},
"Die 7 Tage davor": {CASE_TYPE_HEADERS[case_type]: the_7_days_before},
"Veränderung": {CASE_TYPE_HEADERS[case_type]: percent_change},
},
)
return df
def history(case_type):
cases, stats = filter_data(InsolvencyType.REGULAR)
cases = cases[case_type]
# Bin proceedings by year
by_year = defaultdict(list)
for case in cases:
by_year[case["date"].year].append(case)
# Bin proceedings by year and week
by_week_count = defaultdict(lambda: defaultdict(int))
by_year_and_week_count = defaultdict(lambda: defaultdict(int))
calendar_today = dt.date.today().isocalendar()
year_today = calendar_today[0]
week_today = calendar_today[1]
for case in cases:
# Note: isocalendar week behaves weirdly between years
calendar = case["date"].isocalendar()
year = calendar[0]
week = calendar[1]
# ISO week as string for datawrapper
if year == year_today or year == year_today - 1 and week >= week_today:
subtitle = (
"Unternehmen" if case["courtcase-is-company"] else "Selbstständige"
)
by_week_count[f"{year}W{week}"][
f"{CASE_TYPE_HEADERS[case_type]} ({subtitle})"
] += 1
by_year_and_week_count[year][week] += 1
# Construct dataframes
df_week_year = pd.concat(
{k: pd.Series(v).astype(float) for k, v in by_year_and_week_count.items()},
axis=1,
)
df_week_year.index.name = "Woche"
df_week_year["Durchschnitt 2018/2019"] = (
df_week_year[2018] + df_week_year[2019]
) / 2.0
df_week_year = df_week_year.drop(columns=[2018, 2019])
df_week = pd.concat(
{k: pd.Series(v).astype(float) for k, v in by_week_count.items()},
axis=1,
).T.fillna(0.0)
df_week.index.name = "Woche"
df_week = df_week.reindex(sorted(df_week.columns), axis=1)
return df_week, df_week_year
def districts(case_type):
cases, stats = filter_data(InsolvencyType.REGULAR)
cases = cases[case_type]
# Filter for recent cases
latest_data = max(p["date"] for p in cases)
start_date = latest_data - dt.timedelta(days=30)
last_30_days = [p for p in cases if p["date"] > start_date]
# Group by district name
by_district_name = defaultdict(lambda: defaultdict(int))
for case in last_30_days:
court = next(
residence for residence in case["courtcase-residences"] if in_nrw(residence)
)
district_name = court["geolocation-street"]["street-gemeinde"][
"gemeinde-kreis"
]["kreis-name"]
district_name = normalize.district(district_name)
num_inhabitants = inhabitants[district_name]
by_district_name[district_name][CASE_TYPE_HEADERS[case_type]] += 1
by_district_name[district_name][
f"{CASE_TYPE_HEADERS[case_type]} pro 100.000 Einwohner"
] += (1 / num_inhabitants * 100_000)
# Fill missing districts
for district_name in inhabitants.keys():
if district_name == "Gesamt":
continue
if district_name not in by_district_name:
by_district_name[district_name] = {
CASE_TYPE_HEADERS[case_type]: 0,
f"{CASE_TYPE_HEADERS[case_type]} pro 100.000 Einwohner": 0.0,
}
# Convert to dataframe
df = pd.DataFrame(by_district_name).T
df[CASE_TYPE_HEADERS[case_type]] = df[CASE_TYPE_HEADERS[case_type]].astype(int)
df.index.name = "Name"
df = df.sort_index()
return df
def districts():
cases, stats = filter_data(InsolvencyType.PRIVATE)
proceedings = cases[CaseType.VERFAHRENEROEFFNET]
# Filter for recent proceedings
latest_data = max(p["date"] for p in proceedings)
start_date = latest_data - dt.timedelta(days=30)
last_30_days = [p for p in proceedings if p["date"] > start_date]
# Group by district name
by_district_name = defaultdict(int)
for proceeding in last_30_days:
court = next(residence
for residence in proceeding["courtcase-residences"]
if in_nrw(residence))
district_name = court["geolocation-street"]["street-gemeinde"][
"gemeinde-kreis"]["kreis-name"]
district_name = normalize.district(district_name)
num_inhabitants = inhabitants[district_name]
by_district_name[district_name] += 1 / num_inhabitants * 100_000
# Fill missing districts
for district_name in inhabitants.keys():
if district_name == "Gesamt":
continue
if district_name not in by_district_name:
by_district_name[district_name] = 0.0
# Convert to dataframe
df = pd.DataFrame([by_district_name]).T
df.index.name = "Name"
df = df.rename(columns={0: "pro 100.000 Einwohner"})
return df
def history():
cases, stats = filter_data(InsolvencyType.PRIVATE)
proceedings = cases[CaseType.VERFAHRENEROEFFNET]
# Bin proceedings by year
by_year = defaultdict(list)
for proceeding in proceedings:
by_year[proceeding["date"].year].append(proceeding)
# Bin proceedings by year and week
by_year_and_week = defaultdict(lambda: defaultdict(list))
by_year_and_week_count = defaultdict(lambda: defaultdict(int))
for proceeding in proceedings:
# Note: isocalendar week behaves weirdly between years
calendar = proceeding["date"].isocalendar()
by_year_and_week[calendar[0]][calendar[1]].append(proceeding)
by_year_and_week_count[calendar[0]][calendar[1]] += 1
# Construct dataframe
df_by_week = pd.concat(
{
k: pd.Series(v).astype(float)
for k, v in by_year_and_week_count.items()
},
axis=1,
)
df_by_week.index.name = "Woche"
df_by_week["Durchschnitt 2018/2019"] = (df_by_week[2018] +
df_by_week[2019]) / 2.0
df_by_week = df_by_week.drop(columns=[2018, 2019])
return df_by_week