# -*- coding: utf-8 -*-

# This is a report using the data from IQAASL.
# IQAASL was a project funded by the Swiss Confederation
# It produces a summary of litter survey results for a defined region.
# These charts serve as the models for the development of plagespropres.ch
# The data is gathered by volunteers.
# Please remember all copyrights apply, please give credit when applicable
# The repo is maintained by the community effective January 01, 2022
# There is ample opportunity to contribute, learn and teach
# contact dev@hammerdirt.ch

# Dies ist ein Bericht, der die Daten von IQAASL verwendet.
# IQAASL war ein von der Schweizerischen Eidgenossenschaft finanziertes Projekt.
# Es erstellt eine Zusammenfassung der Ergebnisse der Littering-Umfrage für eine bestimmte Region.
# Diese Grafiken dienten als Vorlage für die Entwicklung von plagespropres.ch.
# Die Daten werden von Freiwilligen gesammelt.
# Bitte denken Sie daran, dass alle Copyrights gelten, bitte geben Sie den Namen an, wenn zutreffend.
# Das Repo wird ab dem 01. Januar 2022 von der Community gepflegt.
# Es gibt reichlich Gelegenheit, etwas beizutragen, zu lernen und zu lehren.
# Kontakt dev@hammerdirt.ch

# Il s'agit d'un rapport utilisant les données de IQAASL.
# IQAASL était un projet financé par la Confédération suisse.
# Il produit un résumé des résultats de l'enquête sur les déchets sauvages pour une région définie.
# Ces tableaux ont servi de modèles pour le développement de plagespropres.ch
# Les données sont recueillies par des bénévoles.
# N'oubliez pas que tous les droits d'auteur s'appliquent, veuillez indiquer le crédit lorsque cela est possible.
# Le dépôt est maintenu par la communauté à partir du 1er janvier 2022.
# Il y a de nombreuses possibilités de contribuer, d'apprendre et d'enseigner.
# contact dev@hammerdirt.ch

# sys, file and nav packages:
import datetime as dt

# math packages:
import pandas as pd
import numpy as np
from scipy import stats
from statsmodels.distributions.empirical_distribution import ECDF

# charting:
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
from matplotlib import ticker
from matplotlib.colors import LinearSegmentedColormap
import seaborn as sns

# home brew utitilties
import resources.chart_kwargs as ck
import resources.sr_ut as sut

# images and display
from IPython.display import Markdown as md

# set some parameters:
start_date = "2020-03-01"
end_date ="2021-05-31"
start_end = [start_date, end_date]
a_fail_rate = 50
unit_label = "p/100m"
a_color = "saddlebrown"

# colors for gradients
cmap2 = ck.cmap2
colors_palette = ck.colors_palette

# set the maps
bassin_map = "resources/maps/thunersee_brienzersee_scaled.jpeg"

# top level aggregation
top = "All survey areas"

# define the feature level and components
this_feature = {'slug':'bt', 'name':"Thunersee / Brienzersee", 'level':'water_name_slug'}
this_level = 'city'
this_bassin = "aare"
bassin_label = "Aare survey area"

lakes_of_interest = [ 'thunersee','brienzersee']
# explanatory variables:
luse_exp = ["% buildings", "% recreation", "% agg", "% woods", "streets km", "intersects"]

# common aggregations
agg_pcs_quantity = {unit_label:"sum", "quantity":"sum"}
agg_pcs_median = {unit_label:"median", "quantity":"sum"}

# aggregation of dimensional data
agg_dims = {"total_w":"sum", "mac_plast_w":"sum", "area":"sum", "length":"sum"}


# columns needed
use_these_cols = ["loc_date" ,
                  "% to buildings",
                  "% to trans", 
                  "% to recreation",
                  "% to agg",
                  "% to woods",
                  "population",
                  this_level,
                  "streets km",
                  "intersects",
                  "length",
                  "groupname",
                  "code"
                 ]

# get your data:
dfBeaches = pd.read_csv("resources/beaches_with_land_use_rates.csv")
dfCodes = pd.read_csv("resources/codes_with_group_names_2015.csv")
dfDims = pd.read_csv("resources/corrected_dims.csv")

# set the index of the beach data to location slug
dfBeaches.set_index("slug", inplace=True)

# make a map to city names
city_map = dfBeaches.city

# map water_name_slug to water_name
wname_wname = dfBeaches[["water_name_slug","water_name"]].reset_index(drop=True).drop_duplicates().set_index("water_name_slug")

dfCodes.set_index("code", inplace=True)

codes_to_change = [
    ["G74", "description", "Insulation foams"],
    ["G940", "description", "Foamed EVA for crafts and sports"],
    ["G96", "description", "Sanitary-pads/tampons, applicators"],
    ["G178", "description", "Metal bottle caps and lids"],
    ["G82", "description", "Expanded foams 2.5cm - 50cm"],
    ["G81", "description", "Expanded foams .5cm - 2.5cm"],
    ["G117", "description", "Expanded foams < 5mm"],
    ["G75", "description", "Plastic/foamed polystyrene 0 - 2.5cm"],
    ["G76", "description", "Plastic/foamed polystyrene 2.5cm - 50cm"],
    ["G24", "description", "Plastic lid rings"],
    ["G33", "description", "Lids for togo drinks plastic"],
    ["G3", "description", "Plastic bags, carier bags"],
    ["G204", "description", "Bricks, pipes not plastic"],
    ["G904", "description", "Plastic fireworks"],
    ["G211", "description", "Swabs, bandaging, medical"],
]

for x in codes_to_change:
    dfCodes = sut.shorten_the_value(x, dfCodes)

# the surveyor designated the object as aluminum instead of metal
dfCodes.loc["G708", "material"] = "Metal"

# make a map to the code descriptions
code_description_map = dfCodes.description

# make a map to the code materials
code_material_map = dfCodes.material

11. Thunersee / Brienzersee

Below: Map of survey locations March 2020 - May 2021. Marker diameter = the mean survey result in pieces of litter per 100 meters (p/100m).

_images/thunersee_brienzersee_3_0.png

11.1. Sample locations

# this is the data before the expanded foams and fragmented plastics are aggregated to Gfrags and Gfoams
before_agg = pd.read_csv("resources/checked_before_agg_sdata_eos_2020_21.csv")

# this is the aggregated survey data that is being used
# a_data is all the data in the survey period
a_data = pd.read_csv(F"resources/checked_sdata_eos_2020_21.csv")
a_data["date"] = pd.to_datetime(a_data.date)

a_data.rename(columns={"% to agg":"% ag", "% to recreation": "% recreation", "% to woods":"% woods", "% to buildings":"% buildings"}, inplace=True)
luse_exp = ["% buildings", "% recreation", "% ag", "% woods", "streets km", "intersects"]

fd = sut.feature_data(a_data, this_feature["level"], these_features=lakes_of_interest)

# cumulative statistics for each code
code_totals = sut.the_aggregated_object_values(fd, agg=agg_pcs_median, description_map=code_description_map, material_map=code_material_map)    

# daily survey totals
dt_all = fd.groupby(["loc_date","location",this_level, "date"], as_index=False).agg(agg_pcs_quantity )

# the materials table
fd_mat_totals = sut.the_ratio_object_to_total(code_totals)

# summary statistics, nsamples, nmunicipalities, names of citys, population
t = sut.make_table_values(fd, col_nunique=["location", "loc_date", "city"], col_sum=["quantity"], col_median=[])

# make a map to the population values for each survey location/city
fd_pop_map = dfBeaches.loc[fd.location.unique()][["city", "population"]].copy()
fd_pop_map.drop_duplicates(inplace=True)

# update t with the population data
t.update(sut.make_table_values(fd_pop_map, col_nunique=["city"], col_sum=["population"], col_median=[]))

# update t with the list of locations from fd
t.update({"locations":fd.location.unique()})

# join the strings into comma separated list
obj_string = "{:,}".format(t["quantity"])
surv_string = "{:,}".format(t["loc_date"])
pop_string = "{:,}".format(int(t["population"]))

# make strings
date_quantity_context = F"For the period between {start_date[:-3]} and {end_date[:-3]}, a total of {obj_string } objects were removed and identified over the course of {surv_string} surveys."
geo_context = F"The {this_feature['name']} results include {t['location']} different locations in {t['city']} different municipalities with a combined population of approximately {pop_string}."
munis_joined = ", ".join(sorted(fd_pop_map["city"]))

# put that all together:
lake_string = F"""
{date_quantity_context} {geo_context }

*{this_feature["name"]} municipalities:*\n\n>{munis_joined}
"""
md(lake_string)

For the period between 2020-03 and 2021-05, a total of 3,764 objects were removed and identified over the course of 36 surveys. The Thunersee / Brienzersee results include 10 different locations in 6 different municipalities with a combined population of approximately 69,120.

Thunersee / Brienzersee municipalities:

Beatenberg, Brienz (BE), Bönigen, Spiez, Thun, Unterseen

11.1.1. Cumulative totals by municipality

dims_parameters = dict(this_level=this_level, 
                       locations=fd.location.unique(), 
                       start_end=start_end, 
                       city_map=city_map, 
                       agg_dims=agg_dims)

dims_table = sut.gather_dimensional_data(dfDims, **dims_parameters)

# a map of total quantity for each component
q_map = fd.groupby(this_level).quantity.sum()

# assgin the quantity and sample numbers to the dims table
for name in dims_table.index:
    dims_table.loc[name, "samples"] = fd[fd[this_level] == name].loc_date.nunique()
    dims_table.loc[name, "quantity"] = q_map[name]

# make a total column for the top feature data, the sum of all the components
dims_table.loc[this_feature["name"]]= dims_table.sum(numeric_only=True, axis=0)

# change the column names to user friendly syle
dims_table.rename(columns=sut.update_dictionary(sut.dims_table_columns), inplace=True)

# format the numercial data
dims_table.sort_values(by=["items"], ascending=False, inplace=True)

# change to formatted strings
dims_table["plastic kg"] = dims_table["plastic kg"]/1000
dims_table[["m²", "meters", "samples", "items"]] = dims_table[["m²", "meters", "samples", "items"]].applymap(lambda x: "{:,}".format(int(x)))
dims_table[["plastic kg", "total kg"]] = dims_table[["plastic kg", "total kg"]].applymap(lambda x: "{:.2f}".format(x))

# figure caption
agg_caption = F"""
*__Below:__ The cumulative weights and measures for {this_feature["name"]} and municipalities*
"""
md(agg_caption)

Below: The cumulative weights and measures for Thunersee / Brienzersee and municipalities

# make table
data = dims_table.reset_index()
colLabels = data.columns

fig, ax = plt.subplots(figsize=(len(colLabels)*2,len(data)*.7))

sut.hide_spines_ticks_grids(ax)
table_one = sut.make_a_table(ax, data.values, colLabels=colLabels, colWidths=[.28, *[.12]*6], a_color="saddlebrown", )
table_one.get_celld()[(0,0)].get_text().set_text(" ")

plt.show()
plt.tight_layout()
plt.close()
_images/thunersee_brienzersee_8_0.png

11.1.2. Distribution of Survey results

# the feature surveys to chart
fd_dindex = dt_all.set_index("date")

# all the other surveys
ots = dict(level_to_exclude=this_feature["level"], components_to_exclude=fd[this_feature["level"]].unique())
dts_date = sut.the_other_surveys(a_data, **ots)

# group the outher surveys by date and total pcs_m
ots_params = dict(agg_this = {unit_label:"sum"}, these_columns = ["loc_date","date"])
dts_date = sut.group_these_columns(dts_date, **ots_params)

# get the monthly or quarterly results for the feature
resample_plot, rate = sut.quarterly_or_monthly_values(fd_dindex , this_feature["name"], vals=unit_label, quarterly=["ticino"])    

# scale the chart as needed to accomodate for extreme values
y_lim = 98
y_limit = np.percentile(dts_date[unit_label], y_lim)

# label for the chart that alerts to the scale
not_included = F"Values greater than {round(y_limit, 1)}{unit_label} not shown."

# figure caption
chart_notes = F"""
*__Left:__ {this_feature['name']}, {start_date[:7]} through {end_date[:7]}, n={t["loc_date"]}. {not_included} __Right:__ {this_feature['name']} empirical cumulative distribution of survey results.*
"""
md(chart_notes )

Left: Thunersee / Brienzersee, 2020-03 through 2021-05, n=36. Values greater than 2375.0p/100m not shown. Right: Thunersee / Brienzersee empirical cumulative distribution of survey results.

# months locator, can be confusing
# https://matplotlib.org/stable/api/dates_api.html
# months = mdates.MonthLocator(interval=1)
months_fmt = mdates.DateFormatter("%b")
days = mdates.DayLocator(interval=7)
sns.set_style("whitegrid")

fig, axs = plt.subplots(1,2, figsize=(10,5))

# the survey totals by day
ax = axs[0]

# feature surveys
sns.scatterplot(data=dts_date, x=dts_date.index, y=unit_label, label=top, color="black", alpha=0.4,  ax=ax)
# all other surveys
sns.scatterplot(data=fd_dindex, x=fd_dindex.index, y=unit_label, label=this_feature["name"], color="red", s=34, ec="white", ax=ax)

# monthly or quaterly plot
sns.lineplot(data=resample_plot, x=resample_plot.index, y=resample_plot, label=F"{this_feature['name']}: {rate} median", color="magenta", ax=ax)

ax.set_ylim(0,y_limit )
ax.set_ylabel(unit_label, **ck.xlab_k14)

ax.set_xlabel("")
ax.xaxis.set_minor_locator(days)
ax.xaxis.set_major_formatter(months_fmt)
ax.legend()

# the cumlative distributions:
axtwo = axs[1]

# the feature of interest
feature_ecd = ECDF(dt_all[unit_label].values)    
sns.lineplot(x=feature_ecd.x, y=feature_ecd.y, color="darkblue", ax=axtwo, label=this_feature["name"])

# the other features
other_features = ECDF(dts_date[unit_label].values)
sns.lineplot(x=other_features.x, y=other_features.y, color="magenta", label=top, linewidth=1, ax=axtwo)

axtwo.set_xlabel(unit_label, **ck.xlab_k14)
axtwo.set_ylabel("Ratio of samples", **ck.xlab_k14)

plt.tight_layout()
plt.show()
_images/thunersee_brienzersee_11_0.png

11.1.3. Summary data and material types

Left: Thunersee / Brienzersee summary of survey totals. Right: Thunersee / Brienzersee material type and percent of total

# get the basic statistics from pd.describe
cs = dt_all[unit_label].describe().round(2)

# change the names
csx = sut.change_series_index_labels(cs, sut.create_summary_table_index(unit_label, lang="EN"))

combined_summary = sut.fmt_combined_summary(csx, nf=[])

fd_mat_totals = sut.fmt_pct_of_total(fd_mat_totals)
fd_mat_totals = sut.make_string_format(fd_mat_totals)

# applly new column names for printing
cols_to_use = {"material":"Material","quantity":"Quantity", "% of total":"% of total"}
fd_mat_t = fd_mat_totals[cols_to_use.keys()].values

# make tables
fig, axs = plt.subplots(1,2, figsize=(8,6))

# summary table
# names for the table columns
a_col = [this_feature["name"], "total"]

axone = axs[0]
sut.hide_spines_ticks_grids(axone)

table_two = sut.make_a_table(axone, combined_summary,  colLabels=a_col, colWidths=[.5,.25,.25],  bbox=[0,0,1,1], **{"loc":"lower center"})
table_two.get_celld()[(0,0)].get_text().set_text(" ")

# material table
axtwo = axs[1]
axtwo.set_xlabel(" ")
sut.hide_spines_ticks_grids(axtwo)

table_three = sut.make_a_table(axtwo, fd_mat_t,  colLabels=list(cols_to_use.values()), colWidths=[.4, .3,.3],  bbox=[0,0,1,1], **{"loc":"lower center"})
table_three.get_celld()[(0,0)].get_text().set_text(" ")

plt.tight_layout()
plt.subplots_adjust(wspace=0.2)
plt.show()
_images/thunersee_brienzersee_14_0.png

11.2. The most common objects

The most common objects are the ten most abundant by quantity AND/OR objects identified in at least 50% of all surveys.

# the top ten by quantity
most_abundant = code_totals.sort_values(by="quantity", ascending=False)[:10]

# the most common
most_common = code_totals[code_totals["fail rate"] >= a_fail_rate].sort_values(by="quantity", ascending=False)

# merge with most_common and drop duplicates
m_common = pd.concat([most_abundant, most_common]).drop_duplicates()

# get percent of total
m_common_percent_of_total = m_common.quantity.sum()/code_totals.quantity.sum()

# figure caption
rb_string = F"""
*__Below:__ {this_feature['name']} most common objects: fail rate >/= {a_fail_rate}%  and/or top ten by quantity. Combined, the most abundant objects represent {int(m_common_percent_of_total*100)}% of all objects found. 
Note : {unit_label} = median survey value.*
"""
md(rb_string)

Below: Thunersee / Brienzersee most common objects: fail rate >/= 50% and/or top ten by quantity. Combined, the most abundant objects represent 72% of all objects found. Note : p/100m = median survey value.

# format values for table
m_common["item"] = m_common.index.map(lambda x: code_description_map.loc[x])
m_common["% of total"] = m_common["% of total"].map(lambda x: F"{x}%")
m_common["quantity"] = m_common.quantity.map(lambda x: "{:,}".format(x))
m_common["fail rate"] = m_common["fail rate"].map(lambda x: F"{x}%")
m_common[unit_label] = m_common[unit_label].map(lambda x: F"{round(x,1)}")

# final table data
cols_to_use = {"item":"Item","quantity":"Quantity", "% of total":"% of total", "fail rate":"Fail rate", unit_label:unit_label}
all_survey_areas = m_common[cols_to_use.keys()].values

colWidths=[.52, .12, .12, .12, .12]
bbox=[0,0,1,1],
kwargs = {"loc":"lower center"}
colLabels=list(cols_to_use.values())
figsize=(len(colLabels)*2,len(all_survey_areas)*.7)

fig, axs = plt.subplots(figsize=figsize)

sut.hide_spines_ticks_grids(axs)

table_four = sut.make_a_table(axs, all_survey_areas, colLabels=colLabels, colWidths=colWidths, a_color="saddlebrown")
table_four.get_celld()[(0,0)].get_text().set_text(" ")
plt.tight_layout()
plt.show()
_images/thunersee_brienzersee_17_0.png

11.2.1. Most common objects results by municipality

Below: Thunersee / Brienzersee most common objects: median p/100m

# aggregated survey totals for the most common codes for all the water features
m_common_st = fd[fd.code.isin(m_common.index)].groupby([this_level, "loc_date","code"], as_index=False).agg(agg_pcs_quantity)
m_common_ft = m_common_st.groupby([this_level, "code"], as_index=False)[unit_label].median()

# map the desctiption to the code
m_common_ft["item"] = m_common_ft.code.map(lambda x: code_description_map.loc[x])

# pivot that
m_c_p = m_common_ft[["item", this_level, unit_label]].pivot(columns=this_level, index="item")

# quash the hierarchal column index
m_c_p.columns = m_c_p.columns.get_level_values(1)

## the aggregated totals for the feature data
c = sut.aggregate_to_group_name(fd[fd.code.isin(m_common.index)], column="code", name=this_feature["name"], val="med")
m_c_p[this_feature["name"]]= sut.change_series_index_labels(c, {x:code_description_map.loc[x] for x in c.index})

# the aggregated totals of the survey area
c = sut.aggregate_to_group_name(a_data[(a_data.river_bassin == this_bassin)&(a_data.code.isin(m_common.index))], column="code", name=top, val="med")
m_c_p[bassin_label] = sut.change_series_index_labels(c, {x:code_description_map.loc[x] for x in c.index})

# the aggregated totals of all the data
c = sut.aggregate_to_group_name(a_data[(a_data.code.isin(m_common.index))], column="code", name=top, val="med")
m_c_p[top] = sut.change_series_index_labels(c, {x:code_description_map.loc[x] for x in c.index})

# chart that
fig, ax  = plt.subplots(figsize=(len(m_c_p.columns)*.9,len(m_c_p)*.9))
axone = ax

sns.heatmap(m_c_p, ax=axone, cmap=cmap2, annot=True, annot_kws={"fontsize":12}, fmt=".1f", square=True, cbar=False, linewidth=.1, linecolor="white")
axone.set_xlabel("")
axone.set_ylabel("")
axone.tick_params(labelsize=14, which="both", axis="x")
axone.tick_params(labelsize=12, which="both", axis="y")

plt.setp(axone.get_xticklabels(), rotation=90)

plt.show()
plt.close()
_images/thunersee_brienzersee_20_0.png

11.2.2. Most common objects monthly average

# collect the survey results of the most common objects
m_common_m = fd[(fd.code.isin(m_common.index))].groupby(["loc_date", "date", "code", "groupname"], as_index=False).agg(agg_pcs_quantity)
m_common_m.set_index("date", inplace=True)

# set the order of the chart, group the codes by groupname columns
an_order = m_common_m.groupby(["code", "groupname"], as_index=False).quantity.sum().sort_values(by="groupname")["code"].values

# a manager dict for the monthly results of each code
mgr = {}

# get the monhtly results for each code:
for a_group in an_order:
    # resample by month
    a_plot = m_common_m[(m_common_m.code==a_group)][unit_label].resample("M").mean().fillna(0)
    this_group = {a_group:a_plot}
    mgr.update(this_group)

monthly_mc = F"""
*__Below:__ {this_feature['name']}, monthly average survey result {unit_label}. Detail of the most common objects*
"""
md(monthly_mc)

Below: Thunersee / Brienzersee, monthly average survey result p/100m. Detail of the most common objects

# convenience function to lable x axis
def new_month(x):
    if x <= 11:
        this_month = x
    else:
        this_month=x-12    
    return this_month
months={
    0:'Jan',
    1:'Feb',
    2:'Mar',
    3:'Apr',
    4:'May',
    5:'Jun',
    6:'Jul',
    7:'Aug',
    8:'Sep',
    9:'Oct',
    10:'Nov',
    11:'Dec'
}

fig, ax = plt.subplots(figsize=(9,8))

# define a bottom
bottom = [0]*len(mgr["G27"])

# the monhtly survey average for all objects and locations
monthly_fd = fd.groupby(["loc_date", "date"], as_index=False).agg(agg_pcs_quantity)
monthly_fd.set_index("date", inplace=True)
m_fd = monthly_fd[unit_label].resample("M").mean().fillna(0)

# define the xaxis
this_x = [i for i,x in  enumerate(m_fd.index)]

# plot the monthly total survey average
ax.bar(this_x, m_fd.to_numpy(), color=a_color, alpha=0.2, linewidth=1, edgecolor="teal", width=1, label="Monthly survey average") 

# plot the monthly survey average of the most common objects
for i, a_group in enumerate(an_order): 
    
    # define the axis
    this_x = [i for i,x in  enumerate(mgr[a_group].index)]
    
    # collect the month
    this_month = [x.month for i,x in enumerate(mgr[a_group].index)]
    
    # if i == 0 laydown the first bars
    if i == 0:
        ax.bar(this_x, mgr[a_group].to_numpy(), label=a_group, color=colors_palette[a_group], linewidth=1, alpha=0.6 ) 
    # else use the previous results to define the bottom
    else:
        bottom += mgr[an_order[i-1]].to_numpy()        
        ax.bar(this_x, mgr[a_group].to_numpy(), bottom=bottom, label=a_group, color=colors_palette[a_group], linewidth=1, alpha=0.8)
        
# collect the handles and labels from the legend
handles, labels = ax.get_legend_handles_labels()

# set the location of the x ticks
ax.xaxis.set_major_locator(ticker.FixedLocator([i for i in np.arange(len(this_x))]))

#label the xticks by month
axisticks = ax.get_xticks()
labelsx = [months[new_month(x-1)] for x in  this_month]
plt.xticks(ticks=axisticks, labels=labelsx)

ax.set_ylabel(unit_label, **ck.xlab_k14)

# make the legend
# swap out codes for descriptions
new_labels = [code_description_map.loc[x] for x in labels[1:]]
new_labels = new_labels[::-1]

# insert a label for the monthly average
new_labels.insert(0,"Monthly average")
handles = [handles[0], *handles[1:][::-1]]
    
plt.legend(handles=handles, labels=new_labels, bbox_to_anchor=(.5, -.05), loc="upper center",  ncol=2, fontsize=14)       
plt.tight_layout()
plt.show()
_images/thunersee_brienzersee_23_0.png

11.3. Utility of the objects found

The utility type is based on the utilization of the object prior to it being discarded or object description if the original use is undetermined. Identified objects are classified into one of 260 predefined categories. The categories are grouped according to utilization or item description.

  • wastewater: items released from water treatment plants includes items likely toilet flushed

  • micro plastics (< 5mm): fragmented plastics and pre-production plastic resins

  • infrastructure: items related to construction and maintenance of buildings, roads and water/power supplies

  • food and drink: all materials related to consuming food and drink

  • agriculture: primarily industrial sheeting i.e., mulch and row covers, greenhouses, soil fumigation, bale wraps. Includes hard plastics for agricultural fencing, flowerpots etc.

  • tobacco: primarily cigarette filters, includes all smoking related material

  • recreation: objects related to sports and leisure i.e., fishing, hunting, hiking etc.

  • packaging non food and drink: packaging material not identified as food, drink nor tobacco related

  • plastic fragments: plastic pieces of undetermined origin or use

  • personal items: accessories, hygiene and clothing related

See the annex for the complete list of objects identified, includes descriptions and group classification. The section Code groups describes each code group in detail and provides a comprehensive list of all objects in a group.



Below: Thunersee / Brienzersee utility of objects found % of total by municipality. Fragmented objects with no clear identification remain classified by size:

# code groups resluts aggregated by survey
groups = ["loc_date","groupname"]
cg_t = fd.groupby([this_level,*groups], as_index=False).agg(agg_pcs_quantity)

# the total per water feature
cg_tq = cg_t.groupby(this_level).quantity.sum()

# get the fail rates for each group per survey
cg_t["fail"]=False
cg_t["fail"] = cg_t.quantity.where(lambda x: x == 0, True)

# aggregate all that for each municipality
agg_this = {unit_label:"median", "quantity":"sum", "fail":"sum", "loc_date":"nunique"} 
cg_t = cg_t.groupby([this_level, "groupname"], as_index=False).agg(agg_this)

# assign survey area total to each record
for a_feature in cg_tq.index:
    cg_t.loc[cg_t[this_level] == a_feature, "f_total"] = cg_tq.loc[a_feature]

# get the percent of total for each group for each survey area
cg_t["pt"] = (cg_t.quantity/cg_t.f_total).round(2)

# pivot that
data_table = cg_t.pivot(columns=this_level, index="groupname", values="pt")

# aggregated values for the lake
data_table[this_feature["name"]]= sut.aggregate_to_group_name(fd, unit_label=unit_label, column="groupname", name=bassin_label, val="pt")

# repeat for the survey area
data_table[bassin_label] = sut.aggregate_to_group_name(a_data[a_data.river_bassin == this_bassin], unit_label=unit_label, column="groupname", name=bassin_label, val="pt")

# repeat for all the data
data_table[top] = sut.aggregate_to_group_name(a_data, unit_label=unit_label, column="groupname", name=top, val="pt")

data = data_table

fig, ax = plt.subplots(figsize=(11,10))

axone = ax
sns.heatmap(data , ax=axone, cmap=cmap2, annot=True, annot_kws={"fontsize":12}, cbar=False, fmt=".0%", linewidth=.1, square=True, linecolor="white")

axone.set_ylabel("")
axone.set_xlabel("")
axone.tick_params(labelsize=14, which="both", axis="both", labeltop=False, labelbottom=True)

plt.setp(axone.get_xticklabels(), rotation=90, fontsize=14)
plt.setp(axone.get_yticklabels(), rotation=0, fontsize=14)

plt.show()
_images/thunersee_brienzersee_26_0.png
cg_medpcm = F"""
<br></br>
*__Below:__ {this_feature['name']} utility of objects found median {unit_label}. Fragmented objects with no clear identification remain classified by size:*
"""
md(cg_medpcm)



Below: Thunersee / Brienzersee utility of objects found median p/100m. Fragmented objects with no clear identification remain classified by size:

# median p/50m of all the water features
data_table = cg_t.pivot(columns=this_level, index="groupname", values=unit_label)

# aggregated values for the lake
data_table[this_feature["name"]]= sut.aggregate_to_group_name(fd, unit_label=unit_label, column="groupname", name=bassin_label, val="med")

# the survey area columns
data_table[bassin_label] = sut.aggregate_to_group_name(a_data[a_data.river_bassin == this_bassin], unit_label=unit_label, column="groupname", name=bassin_label, val="med")

# column for all the surveys
data_table[top] = sut.aggregate_to_group_name(a_data, unit_label=unit_label, column="groupname", name=top, val="med")

# merge with data_table
data = data_table

fig, ax = plt.subplots(figsize=(11,10))

axone = ax
sns.heatmap(data , ax=axone, cmap=cmap2, annot=True, annot_kws={"fontsize":12}, fmt="g", cbar=False, linewidth=.1, square=True, linecolor="white")

axone.set_xlabel("")
axone.set_ylabel("")
axone.tick_params(labelsize=14, which="both", axis="both", labeltop=False, labelbottom=True)

plt.setp(axone.get_xticklabels(), rotation=90, fontsize=14)
plt.setp(axone.get_yticklabels(), rotation=0, fontsize=14)

plt.show()
_images/thunersee_brienzersee_28_0.png

11.4. Annex

11.4.1. Fragmented foams and plastics by size

The table below contains the “Gfoam” and “Gfrags” components grouped for analysis. Objects labeled expanded foams are grouped as Gfoam and includes all expanded polystyrene foamed plastics > 0.5 cm. Plastic pieces and objects made of combined plastic and foamed plastic materials > 0.5 cm. are grouped for analysis as Gfrags.

Below: Thunersee / Brienzersee fragmented foams and plastics by size group.

# collect the data before aggregating foams for all locations in the survey area
# the codes for the foams
some_foams = ["G81", "G82", "G83", "G74"]

# the codes for the fragmented plastics
some_frag_plas = list(before_agg[before_agg.groupname == "plastic pieces"].code.unique())

# aggregate all the codes by loc_date and get the total quantity and the median pcs/m
fd_frags_foams = before_agg[(before_agg.code.isin([*some_frag_plas, *some_foams]))&(before_agg.location.isin(t["locations"]))].groupby(["loc_date","code"], as_index=False).agg(agg_pcs_quantity)
fd_frags_foams = fd_frags_foams.groupby("code").agg({unit_label:"median", "quantity":"sum"})

# add code description and format for printing
fd_frags_foams["item"] = fd_frags_foams.index.map(lambda x: code_description_map.loc[x])
fd_frags_foams["% of total"] = (fd_frags_foams.quantity/fd.quantity.sum()*100).round(2)
fd_frags_foams["% of total"] = fd_frags_foams["% of total"].map(lambda x: F"{x}%")
fd_frags_foams["quantity"] = fd_frags_foams["quantity"].map(lambda x: F"{x:,}")

# table data
data = fd_frags_foams[["item", unit_label, "quantity", "% of total"]]

fig, axs = plt.subplots(figsize=(11,len(data)*.7))
sut.hide_spines_ticks_grids(axs)

a_table = sut.make_a_table(axs, data.values,  colLabels=data.columns, colWidths=[.6, .13, .13, .13])
a_table.get_celld()[(0,0)].get_text().set_text(" ")


plt.show()
plt.tight_layout()
plt.close()
_images/thunersee_brienzersee_31_0.png

11.4.2. The survey locations

# display the survey locations

disp_columns = ["latitude", "longitude", "city"]
disp_beaches = dfBeaches.loc[t["locations"]][disp_columns]
disp_beaches.reset_index(inplace=True)
disp_beaches.rename(columns={"slug":"location"}, inplace=True)
disp_beaches.set_index("location", inplace=True, drop=True)

disp_beaches
latitude longitude city
location
weissenau-neuhaus 46.676583 7.817528 Unterseen
thun-strandbad 46.739939 7.633520 Thun
oben-am-see 46.744451 8.049921 Brienz (BE)
thunersee_spiez_meierd_1 46.704437 7.657882 Spiez
delta-park 46.720078 7.635304 Spiez
sundbach-strand 46.684386 7.794768 Beatenberg
wycheley 46.740370 8.048640 Brienz (BE)
camping-gwatt-strand 46.727140 7.629620 Thun
augustmutzenbergstrandweg 46.686640 7.689760 Spiez
hafeli 46.690283 7.898592 Bönigen

11.4.3. Inventory of items

pd.set_option("display.max_rows", None)

complete_inventory = code_totals[code_totals.quantity>0][["item", "groupname", "quantity","% of total","fail rate"]]
complete_inventory.sort_values(by="quantity", ascending=False)
item groupname quantity % of total fail rate
code
G27 Cigarette filters tobacco 748 19.87 88
Gfrags Fragmented plastics plastic pieces 515 13.68 91
G67 Industrial sheeting agriculture 352 9.35 83
Gfoam Expanded polystyrene infrastructure 281 7.47 83
G30 Food wrappers; candy, snacks food and drink 228 6.06 77
G74 Insulation foams infrastructure 158 4.20 66
G200 Glass drink bottles, pieces food and drink 141 3.75 72
G112 Industrial pellets (nurdles) micro plastics (< 5mm) 115 3.06 11
G941 Packaging films nonfood or unknown packaging non food 83 2.21 44
G89 Plastic construction waste infrastructure 65 1.73 50
G904 Plastic fireworks recreation 39 1.04 38
G21 Drink lids food and drink 38 1.01 38
G117 Expanded foams < 5mm micro plastics (< 5mm) 36 0.96 33
G156 Paper fragments packaging non food 35 0.93 22
G95 Cotton bud/swab sticks waste water 35 0.93 52
G106 Plastic fragments angular <5mm micro plastics (< 5mm) 34 0.90 25
G944 Pellet mass from injection molding unclassified 34 0.90 8
G23 Lids unidentified packaging non food 30 0.80 30
G24 Plastic lid rings food and drink 28 0.74 36
G213 Paraffin wax recreation 27 0.72 33
G177 Foil wrappers, aluminum foil food and drink 26 0.69 36
G936 Sheeting ag. greenhouse film agriculture 26 0.69 22
G50 String < 1cm recreation 25 0.66 33
G922 Labels, bar codes packaging non food 24 0.64 33
G153 Cups, food containers, wrappers (paper) food and drink 21 0.56 11
G186 Industrial scrap infrastructure 20 0.53 25
G25 Tobacco; plastic packaging, containers tobacco 19 0.50 27
G201 Jars, includes pieces food and drink 18 0.48 16
G10 Food containers single use foamed or plastic food and drink 18 0.48 27
G211 Swabs, bandaging, medical personal items 17 0.45 27
G35 Straws and stirrers food and drink 17 0.45 30
G32 Toys and party favors recreation 16 0.43 25
G87 Tape, masking/duct/packing infrastructure 16 0.43 19
G908 Tape; electrical, insulating infrastructure 16 0.43 22
G91 Biomass holder waste water 16 0.43 27
G31 Lollypop sticks food and drink 16 0.43 33
G943 Fencing agriculture, plastic agriculture 14 0.37 8
G208 Glass or ceramic fragments > 2.5 cm unclassified 14 0.37 11
G204 Bricks, pipes not plastic infrastructure 14 0.37 16
G178 Metal bottle caps and lids food and drink 14 0.37 25
G66 Straps/bands; hard, plastic package fastener infrastructure 14 0.37 33
G927 String trimmer line, used to cut grass, weeds,... infrastructure 13 0.35 16
G70 Shotgun cartridges recreation 13 0.35 22
G210 Other glass/ceramic unclassified 12 0.32 5
G90 Plastic flower pots agriculture 12 0.32 19
G942 Plastic shavings from lathes, CNC machining unclassified 11 0.29 11
G3 Plastic bags, carier bags packaging non food 11 0.29 19
G131 Rubber bands personal items 10 0.27 19
G923 Tissue, toilet paper, napkins, paper towels personal items 10 0.27 19
G149 Paper packaging packaging non food 10 0.27 11
G203 Tableware ceramic or glass, cups, plates, pieces food and drink 10 0.27 11
G98 Diapers - wipes waste water 10 0.27 19
G152 Cigarette boxes, tobacco related paper/cardboard tobacco 10 0.27 16
G905 Hair clip, hair ties, personal accessories pl... personal items 9 0.24 16
G65 Buckets agriculture 8 0.21 8
G100 Medical; containers/tubes/ packaging waste water 8 0.21 16
G22 Lids for chemicals, detergents (non-food) infrastructure 8 0.21 8
G48 Rope, synthetic recreation 8 0.21 13
G33 Lids for togo drinks plastic food and drink 8 0.21 16
G928 Ribbons and bows personal items 8 0.21 8
G159 Corks food and drink 7 0.19 16
G175 Cans, beverage food and drink 7 0.19 13
G73 Foamed items & pieces (non packaging/insulatio... recreation 6 0.16 13
G198 Other metal pieces < 50cm infrastructure 6 0.16 16
G931 Tape-caution for barrier, police, construction... infrastructure 5 0.13 5
G125 Balloons and balloon sticks recreation 5 0.13 11
G170 Wood (processed) agriculture 5 0.13 13
G34 Cutlery, plates and trays food and drink 5 0.13 13
G191 Wire and mesh agriculture 5 0.13 11
G940 Foamed EVA for crafts and sports recreation 5 0.13 5
G155 Fireworks paper tubes and fragments recreation 5 0.13 13
G4 Small plastic bags; freezer, zip-lock etc. packaging non food 5 0.13 11
G148 Cardboard (boxes and fragments) packaging non food 5 0.13 8
G917 Terracotta balls unclassified 5 0.13 11
G925 Packets: desiccant/ moisture absorbers, plasti... packaging non food 4 0.11 5
G59 Fishing line monofilament (angling) recreation 4 0.11 8
G194 Cables, metal wire(s) often inside rubber or p... infrastructure 4 0.11 8
G96 Sanitary-pads/tampons, applicators waste water 4 0.11 11
G176 Cans, food food and drink 3 0.08 5
G913 Pacifier personal items 3 0.08 8
G933 Bags, cases for accessories; glasses, electron... personal items 3 0.08 8
G37 Mesh bags personal items 3 0.08 8
G6 Bottles and containers, plastic non food/drink packaging non food 3 0.08 2
G914 Paperclips, clothespins, plastic utility items personal items 3 0.08 8
G921 Ceramic tile and pieces infrastructure 3 0.08 8
G134 Other rubber unclassified 3 0.08 5
G158 Other paper items packaging non food 3 0.08 8
G146 Paper, cardboard packaging non food 3 0.08 2
G165 Ice cream sticks, toothpicks, chopsticks food and drink 3 0.08 8
G103 Plastic fragments rounded <5mm micro plastics (< 5mm) 2 0.05 5
G918 Safety pins, paper clips, small metal utility ... personal items 2 0.05 5
G101 Dog feces bag personal items 2 0.05 5
G124 Other plastic or foam products unclassified 2 0.05 5
G133 Condoms incl. packaging waste water 2 0.05 5
G2 Bags packaging non food 2 0.05 2
G174 Aerosol spray cans infrastructure 2 0.05 5
G49 Rope > 1cm recreation 2 0.05 5
G26 Cigarette lighters tobacco 2 0.05 5
G901 Mask medical, synthetic personal items 1 0.03 2
G938 Toothpicks, dental floss plastic food and drink 1 0.03 2
G939 Flowers, plants plastic personal items 1 0.03 2
G20 Caps and lids packaging non food 1 0.03 2
G111 Spheruloid pellets < 5mm micro plastics (< 5mm) 1 0.03 2
G11 Cosmetics for the beach, e.g. sunblock recreation 1 0.03 2
G197 Other metal infrastructure 1 0.03 2
G202 Light bulbs unclassified 1 0.03 2
G190 Paint cans infrastructure 1 0.03 2
G182 Fishing; hooks, weights, lures, sinkers etc. recreation 1 0.03 2
G97 Toilet fresheners waste water 1 0.03 2
G161 Processed timber agriculture 1 0.03 2
G114 Films <5mm micro plastics (< 5mm) 1 0.03 2
G930 Foam earplugs personal items 1 0.03 2
G214 Oil/tar infrastructure 1 0.03 2
G129 Inner tubes and rubber sheets unclassified 1 0.03 2
G93 Cable ties; steggel, zip, zap straps infrastructure 1 0.03 2
G138 Shoes and sandals personal items 1 0.03 2
G926 Chewing gum, often contains plastics food and drink 1 0.03 2
G135 Clothes, footware, headware, gloves personal items 1 0.03 2
G29 Combs, brushes and sunglasses personal items 1 0.03 2
G12 Cosmetics, non-beach use personal care containers personal items 1 0.03 2
G36 Bags/sacks heavy duty plastic for 25 Kg or mor... agriculture 1 0.03 2
G919 Nails, screws, bolts etc. infrastructure 1 0.03 2
G43 Tags fishing or industry (security tags, seals) recreation 1 0.03 2
G167 Matches or fireworks recreation 1 0.03 2
G64 Fenders unclassified 1 0.03 2
G71 Shoes sandals personal items 1 0.03 2
G28 Pens, lids, mechanical pencils etc. personal items 1 0.03 2