Py.Cafe

import pandas as pd
from dash import Dash, html, dcc, Input, Output, callback
import dash_bootstrap_components as dbc
import numpy as np
import plotly.express as px


df = pd.read_csv("https://raw.githubusercontent.com/plotly/Figure-Friday/refs/heads/main/2025/week-2/samples.tsv",  sep='\t')

#replace blinded name "Tablet" with "Tablets" in blinded_name
#replace blinded name "Protein Bar" with "Protein bar" in blinded_name

df.loc[df['blinded_name'] == 'Tablet', 'blinded_name'] = 'Tablets'
df.loc[df['blinded_name'] == 'Protein Bar', 'blinded_name'] = 'Protein bar'
#may some grouping/dropdown thing
df_chemicals = pd.read_csv("chemicals.csv")



#distinct blind_names = 117, samples = 618
#blind name in percentile, like fish percentile scores for some of the ng/g stats
#percentile nanogram per gram columns are:
percentile_cols = [col for col in df.columns if 'percentile' in col and '_g' in col and 'equivalents' not in col]

#start new df with non-percentile columns to use
df_data = df[['id','product_id', 'product', 'blinded_name']].copy()






#append percentile for ng/gram to dataframe
for x in percentile_cols:
    df_data[x] = df[x].apply(lambda x: x if x in ['<LOQ', 'NO RESULT']
                             else ('Q4' if int(x) > 75 
                             else ('Q3' if int(x) >50
                             else ('Q2' if int(x) >25 
                             else ('Q1' if int(x) >= 0 
                             else 'Error'
                                   )))))


#convert values in columns to something to use in a visual

def convert_percentiles_into_dict(a):
    
    #input column, output dict with countvalues
    #inititialize empty dict

    count_sample_output = {'<LOQ': 0, 'Q1': 0, 'Q2': 0,'Q3':0, 'Q4':0}
    #count
    count_sample_output['Q1'] = int(a.value_counts().get('Q1', 0))
    count_sample_output['Q2'] = int(a.value_counts().get('Q2', 0))
    count_sample_output['Q3'] = int(a.value_counts().get('Q3', 0))
    count_sample_output['Q4'] = int(a.value_counts().get('Q4', 0))
    count_sample_output['<LOQ'] = int(a.value_counts().get('<LOQ', 0))
    
    
    return count_sample_output

def create_productlist_fromcol(a):
    print(a.unique())
    productlist = list(a.unique())
    
    return productlist



#number of samples per blinded_name and number of products tested and number of samples tested
#dict with summary of LOQ and P values

df_grouped_blinded_names = df_data.groupby(['blinded_name']).agg(
        
    number_of_samples = pd.NamedAgg('id', 'count'),
    number_of_different_products=pd.NamedAgg(column="product_id", aggfunc=lambda x: x.nunique()),
    productlist = pd.NamedAgg(column="product", aggfunc=lambda x: ', '.join(x.unique())),
    DEHP_results = pd.NamedAgg(column="DEHP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)), 
    DBP_results = pd.NamedAgg(column="DBP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    BBP_results = pd.NamedAgg(column="BBP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DINP_results = pd.NamedAgg(column="DINP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DIDP_results = pd.NamedAgg(column="DIDP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DEP_results = pd.NamedAgg(column="DEP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DMP_results = pd.NamedAgg(column="DMP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DIBP_results = pd.NamedAgg(column="DIBP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DNHP_results = pd.NamedAgg(column="DNHP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DCHP_results = pd.NamedAgg(column="DCHP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DNOP_results = pd.NamedAgg(column="DNOP_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    BPA_results = pd.NamedAgg(column="BPA_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    BPS_results = pd.NamedAgg(column="BPS_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    BPF_results = pd.NamedAgg(column="BPF_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DEHT_results = pd.NamedAgg(column="DEHT_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DEHA_results = pd.NamedAgg(column="DEHA_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DINCH_results = pd.NamedAgg(column="DINCH_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    DIDA_results = pd.NamedAgg(column="DIDA_percentile_ng_g", aggfunc=lambda x: convert_percentiles_into_dict(x)),
    
    
    ).reset_index()


#create categorical barchart (categories is LOQ, P's)



def create_productlist_card(selected_food_group):
    
    
    productlist = str(df_grouped_blinded_names.loc[df_grouped_blinded_names['blinded_name']==selected_food_group]['productlist'].item()).split(',')  
   #print(productlist)
    
    listChildren = []
    
    for x in productlist:
        listChildren.append(dbc.ListGroupItem(x))
        

    
    productlist_card = dbc.Col(
       dbc.Card([   html.P('Disclaimer: this app offers a way to look at the data collected and analysed by plasticlist.org. Visuals are based on percentiles nanogram/gram. ChatGPT offered me 5 free tweet length descriptions for a chemical.',style={"fontWeight":"bold","fontSize":"11px"}), 
                    html.P('Do not draw any conclusions from here, go to plasticlist.org to read all about the research.',style={"fontWeight":"bold","fontSize":"11px"}),
                    html.A("Plasticlist.org", href='https://plastlist.org', target="_blank",style={"fontWeight":"bold","fontSize":"11px"}),
                    html.P("PlasticList. 'Data on Plastic Chemicals in Bay Area Foods'. plasticlist.org. Accessed Jan 10, 2025.",style={"fontWeight":"bold","fontSize":"11px"}),
                    html.H3("Productlist:", style={"marginTop":"1rem"}),
                    dbc.ListGroup(id = "productlist", children = listChildren, flush=True, style={"fontSize":"10px"})],
                    ), className='col-lg-3 col-md-6 col-sm-12'
       )
    
    
    return productlist_card
    
    

def create_chemical_card(dfin, food_group, chemical):
    
    #get description if available
    desc=df_chemicals.loc[df_chemicals['Chemical'] == chemical]['Description'].item() 
    if len(str(desc)) == 0:
        desc = ''
    
    
    #locate datadict, when are you going to remember to add .item()
    colname = chemical + '_results'
    d=dfin.loc[dfin['blinded_name'] == food_group][colname].item()   
    #colorset for results
    color_map = {'<LOQ':'darkgrey','Q1':'#F4E5CC','Q2':'#eacb99','Q3': '#e0b166','Q4':'#d69732'}
    #extract x en y from dict
    x_vals = list(d.keys())
    y_vals = list(d.values()) 
    #ymax equals number of samples for blinded name
    y_max=dfin.loc[dfin['blinded_name'] == food_group]['number_of_samples'].item()  
    
    fig = px.bar(x=x_vals, y=y_vals, color=x_vals, color_discrete_map = color_map)
    fig.update_layout(showlegend=False,
                      margin=dict(l=10, r=10, t=10, b=10),
                      xaxis=dict(
    title=dict(
        text="Result"
    )
),
yaxis=dict(
    title=None
),)
    fig.update_yaxes(range=[0,y_max+1])

    
    card_chemical = dbc.Col(
       dbc.Card([   html.H3(chemical),
                    html.P(desc, style={"font-size": "11px"}),
                    dcc.Graph(id= str(chemical)+"_plot", figure = fig, style={'height':'350px'})],
                    ), className='col-lg-3 col-md-6 col-sm-12'
       )
  
    return card_chemical


def bar_chart_grid(dfin, selected_chemical_group, df_chemicals,food_group):
    
    #create list with chemicals in chemical group
    
    chemicals_list = sorted(list(df_chemicals.loc[df_chemicals['Chemical Group'] == selected_chemical_group]['Chemical']))
    
    card_list=[create_productlist_card(food_group)]
    
    for x in chemicals_list:
        
    
        card_list.append(create_chemical_card(dfin, food_group, x))
        
        
        
    
   # print(card_list) 
    return card_list



dbc_css = "https://cdn.jsdelivr.net/gh/AnnMarieW/dash-bootstrap-templates/dbc.min.css"
app = Dash(__name__, external_stylesheets=[dbc.themes.SANDSTONE, dbc_css])



app.layout = dbc.Container(
    [dbc.Row([
        html.Div([
            html.H2(id="h2header"),
            
            
            ], className='col-md-7'),
        html.Div([ dcc.Dropdown(id='dropdown_food',  options=[
        {'label': i, 'value': i} for i in df_grouped_blinded_names['blinded_name'].unique()
    ], multi=False, value='Almond milk', placeholder='Filter food category...')], className = "col-md-3"),
        html.Div([dcc.Dropdown(id='dropdown_chemicalgroups', options=[
       {'label': i, 'value': i} for i in df_chemicals['Chemical Group'].unique()
   ], multi=False, value='Phthalates' , placeholder='Filter chemical group...')],  className = 'col-md-2')
        ], style={"position":"fixed", "top": "0", "left": "0", "width": "100%","height":"100px",
                                           "backgroundColor":"lightblue","zIndex":"1000", "margin":"0",
                                           "padding": "1rem", "display":"flex", "alignItems":"center"}),   
     
     
     
     dbc.Row([
        dbc.Col([
            
            html.Div(id="bargrid", style={"display": "flex", "flexWrap":"wrap"} , children=bar_chart_grid(df_grouped_blinded_names, 'Phthalate substitutes', df_chemicals,'Baby formula'))

            #dcc.Graph(id="scatter-plot", figure = create_bar_chart(df_grouped_blinded_names, 'Baby formula')),
            
           
            
            ], className = 'col-md-12')

          
        ], style={"marginTop": "120px"}),
        #add to row link to https://www.plasticlist.org/
        dbc.Row([html.P('Data: PlasticList. "Data on Plastic Chemicals in Bay Area Foods". plasticlist.org. Accessed Jan 10, 2025.')])
], style={"marginTop": "2rem"})


@app.callback( Output('bargrid', 'children'),
              Output('h2header', 'children'),
    #          Output('productlist', 'children'),
              Input(component_id='dropdown_food', component_property='value'),
              Input(component_id='dropdown_chemicalgroups', component_property='value')
              
              )





def update_job_data(selected_food_group, selected_chemical_group):
    
    
    newgrid =bar_chart_grid(df_grouped_blinded_names, selected_chemical_group, df_chemicals, selected_food_group)
    
    samples = str(df_grouped_blinded_names.loc[df_grouped_blinded_names['blinded_name']==selected_food_group]['number_of_samples'].item())
    products = str(df_grouped_blinded_names.loc[df_grouped_blinded_names['blinded_name']==selected_food_group]['number_of_different_products'].item()) 
    h2 = f"{selected_food_group}: {samples} sample(s), {products} different product(s)"   
   
    
    return newgrid,h2





app.run_server(debug=True)