streamlit_app.py

import backend as be
import ui_helpers as uih
import streamlit as st
import matplotlib
import plotly.express as px

st.header('How did your County Change During Covid?')

# Let the user select a (state, county, demographic) combination to get data on
# State and County dropdowns appear side by side
state_col, county_col = st.columns(2)
with state_col:
    state_name = st.selectbox("State:", be.get_state_names(), index=4) # 4 = California
    county_name_index = uih.get_county_name_index(state_name)
with county_col:
    county_name = st.selectbox("County:", be.get_county_names(state_name), index=county_name_index)

# Demographic statistic dropdown appears below
var = st.selectbox("Demographic:", be.get_unique_census_labels()) # Something like "Total Population"

# Now display the data the user requested 
tab1, tab2, tab3, tab4 = st.tabs(["📈 Details", "🥇 Rankings", "🗺️ Map", "ℹ️ About"])

# Tab 1: Time series data on selected county / demographic combination
with tab1:
    st.write(f"All data for **{county_name}, {state_name}** for **{var}**.")
    df = be.get_census_data(state_name, county_name, var)

    col1, col2 = st.columns(2)
    with col1:
        # Line graph of raw data. Set y-axis formatter to use commas
        fig = df.plot(x='YEAR', y=var, style='-o').figure
        fig.gca().get_yaxis().set_major_formatter(
            matplotlib.ticker.StrMethodFormatter('{x:,.0f}')
        )
        st.pyplot(fig)
    with col2:
        # Bar plot showing % change
        df['Percent Change'] = df[var].pct_change() * 100
        st.pyplot(df.plot(kind='bar', x='YEAR', y='Percent Change').figure)

# Tab 2: Ranking of all counties for that demographic (2019-2021)
with tab2:
    ranking_df = be.get_ranking_df(var)
    ranking_text = be.get_ranking_text(state_name, county_name, var, ranking_df)

    st.write(ranking_text)
    # The styling here are things like the gradient on the "Percent Change" column
    ranking_df = ranking_df.style.pipe(uih.apply_styles, state_name, county_name)
    st.dataframe(ranking_df)

# Tab 3: Choropleth map
with tab3:
    st.write("Data is provided only for counties with a population of at least 65,000.")          
    fig = px.choropleth(be.get_mapping_df(var), geojson=be.county_map, locations='FIPS', color='Quartile',
                        color_discrete_sequence = ['#ffffcc','#a1dab4','#41b6c4','#225ea8'],
                        scope="usa",
                        hover_name='County',
                        hover_data={'FIPS': False, 'Percent Change': True},
                        labels={'Quartile':'Percent Change', 'FIPS': 'NAME'})
    st.plotly_chart(fig)

# Tab 4: Info about the data / app
with tab4:
    text = open('about.md').read()
    st.write(text)

st.write("Created by [Ari Lamstein](https://www.arilamstein.com). View the code [here](https://github.com/arilamstein/censusdis-streamlit).")