src/streamlit_app.py

# streamlit_app.py
"""
Chicago Parks in Motion — Streamlit app
Author: Juhi Khare (jkhare2), Alisha Rawat (alishar4), Sutthana Koo-Anupong (sk188)
Primary dataset: Chicago Park District Activities
Data source (CSV endpoint): https://data.cityofchicago.org/resource/tn7v-6rnw.csv
"""

import streamlit as st
import pandas as pd
import numpy as np
import plotly.express as px


st.set_page_config(page_title="Chicago Parks in Motion", layout="wide")

# -------------------------
# Helper: Load & preprocess
# -------------------------
@st.cache_data(ttl=3600)
def load_data():
    csv_url = "https://data.cityofchicago.org/resource/tn7v-6rnw.csv"
    try:
        df = pd.read_csv(csv_url, dtype=str)
    except Exception as e:
        st.error("Could not load dataset from the City of Chicago portal.")
        raise e

    df.columns = [c.strip() for c in df.columns]

    if "fee" in df.columns:
        df["fee"] = pd.to_numeric(df["fee"], errors="coerce")

    # Extract lat/lon
    def extract_latlon(val):
        if pd.isna(val):
            return (np.nan, np.nan)
        sval = str(val)
        if "POINT" in sval:
            try:
                inside = sval.split("(", 1)[1].rstrip(")")
                lon, lat = map(float, inside.split())
                return lat, lon
            except:
                return (np.nan, np.nan)
        return (np.nan, np.nan)

    if "location" in df.columns:
        latlon = df["location"].map(extract_latlon)
        df["latitude"] = latlon.map(lambda x: x[0])
        df["longitude"] = latlon.map(lambda x: x[1])
    else:
        df["latitude"] = np.nan
        df["longitude"] = np.nan

    # Dates
    for c in ["start_date", "end_date"]:
        if c in df.columns:
            df[c] = pd.to_datetime(df[c], errors="coerce")

    # Activity type clean
    if "activity_type" in df.columns:
        df["activity_type_clean"] = df["activity_type"].str.title().fillna("Unknown")
    else:
        df["activity_type_clean"] = "Unknown"

    # Park name
    possible_names = ["park_name", "park", "location_facility", "location_name", "site_name"]
    park_col = next((col for col in possible_names if col in df.columns), None)
    if park_col:
        df["park_name"] = df[park_col].astype(str).replace(["", "nan", "None"], "Unknown Park")
    else:
        df["park_name"] = "Unknown Park"

    return df


df = load_data()

# -------------------------
# Title
# -------------------------
st.title("Chicago Parks in Motion: How Our City Plays")
st.markdown("**Authors:** Juhi Khare (jkhare2), Alisha Rawat (alishar4), Sutthana Koo-Anupong (sk188)")

# -------------------------
# Sidebar filters
# -------------------------
st.sidebar.header("Filters & Settings")

categories = sorted(df["activity_type_clean"].dropna().unique())
chosen_category = st.sidebar.selectbox("Activity category", ["All"] + categories)

# Season detection
def season_from_date(dt):
    if pd.isna(dt): return "Unknown"
    m = dt.month
    if m in [12,1,2]: return "Winter"
    if m in [3,4,5]: return "Spring"
    if m in [6,7,8]: return "Summer"
    return "Fall"

df["season"] = df["start_date"].map(season_from_date)
seasons = sorted(df["season"].unique())
chosen_season = st.sidebar.selectbox("Season", ["All"] + seasons)

if "fee" in df.columns:
    max_fee = float(df["fee"].fillna(0).max())
    fee_limit = st.sidebar.slider("Maximum fee (USD)", 0.0, max_fee, max_fee)
else:
    fee_limit = None

park_search = st.sidebar.text_input("Search park name (partial)")

# Accessibility hint
st.sidebar.caption("Filters help beginners explore the dataset easily without technical skills.")

# -------------------------
# Filtering logic
# -------------------------
filtered = df.copy()
if chosen_category != "All":
    filtered = filtered[filtered["activity_type_clean"] == chosen_category]
if chosen_season != "All":
    filtered = filtered[filtered["season"] == chosen_season]
if fee_limit is not None:
    filtered = filtered[filtered["fee"].fillna(0) <= fee_limit]
if park_search:
    filtered = filtered[filtered["park_name"].str.contains(park_search, case=False)]

st.sidebar.write(f"Programs shown: **{len(filtered):,}**")

# -------------------------
# CENTRAL VISUALIZATION
# -------------------------
st.header("Central Interactive Visualization — Programs by Park")

view = st.radio("Choose a view:", ["Map (recommended)", "Bar chart"], horizontal=True)

if view.startswith("Map"):
    # Aggregate for map
    agg = (
        filtered.groupby(["park_name", "latitude", "longitude"], dropna=True)
        .size().reset_index(name="count")
    )

    if agg.dropna().shape[0] > 0:
        fig_map = px.scatter_mapbox(
            agg,
            lat="latitude",
            lon="longitude",
            size="count",
            color="count",
            color_continuous_scale="Bluered",
            size_max=28,
            zoom=10,
            hover_name="park_name",
            hover_data={"count": True},
            height=600,
        )
        fig_map.update_layout(mapbox_style="open-street-map", margin=dict(l=0,r=0,b=0,t=0))
        st.plotly_chart(fig_map, use_container_width=True)
    else:
        st.warning("No geographic coordinates available for this filtered view.")
else:
    agg = filtered.groupby("park_name").size().reset_index(name="count")
    agg = agg.sort_values("count", ascending=False).head(20)

    fig_bar = px.bar(
        agg,
        x="count",
        y="park_name",
        orientation="h",
        color="count",
        color_continuous_scale="Cividis",
        height=600,
    )
    fig_bar.update_layout(yaxis={'categoryorder':'total ascending'})
    st.plotly_chart(fig_bar, use_container_width=True)

# Explanation under central viz
st.markdown("""
**What this visualization shows:**  
This is our main visualization because it helps readers understand where activities are happening across Chicago’s parks.  
The map shows each park as a circle, where larger and darker circles represent locations with more programs.  
This makes it easy to see which areas are activity hubs and which are quieter. The filters allow anyone to explore patterns by season, 
category, price, or park—without needing technical experience.
""")

# -------------------------
# CONTEXTUAL VISUALIZATION 1
# -------------------------
st.header("Contextual Visualization 1 — Activity Category Breakdown")

cat_counts = df["activity_type_clean"].value_counts().reset_index()
cat_counts.columns = ["activity_type", "count"]

fig_cat = px.pie(
    cat_counts,
    names="activity_type",
    values="count",
    hole=0.35,
    color_discrete_sequence=px.colors.sequential.RdBu
)
st.plotly_chart(fig_cat, use_container_width=True)

st.markdown("""
**Why this matters:**  
This chart shows what kinds of activities Chicago parks offer most often—such as sports, aquatics, arts, or youth programs.  
It helps readers understand the variety of programs available across the city.  
Using a simple color palette keeps the chart readable for people who may not be familiar with data visualization.
""")

# -------------------------
# CONTEXTUAL VISUALIZATION 2
# -------------------------
st.header("Contextual Visualization 2 — Programs by Season")

season_counts = df["season"].value_counts().reset_index()
season_counts.columns = ["Season", "Program Count"]

fig_season = px.bar(
    season_counts,
    x="Season",
    y="Program Count",
    color="Program Count",
    color_continuous_scale="Tealgrn",
    text="Program Count",
    height=500,
)
fig_season.update_traces(textposition="outside")

st.plotly_chart(fig_season, use_container_width=True)

st.markdown("""
**Why this is helpful:**  
This chart shows when programs are most active throughout the year.  
Comparing seasons helps readers see whether summer is the busiest time, or whether activities are spread evenly.  
This makes it easier for residents and planners to understand how weather, school schedules, and community needs 
shape the timing of park programs.
""")

# -------------------------
# FINAL 3-PARAGRAPH EXPLANATION (as provided by you, unchanged)
# -------------------------
st.header("📝 What this data story is showing")

st.markdown("""
Chicago’s parks offer many kinds of activities for people of all ages. These include sports, arts, fitness classes, youth programs, and seasonal events. Each row in this dataset represents one program offered at a park. Our main interactive map helps readers quickly see which parks offer the most activities. Bigger or darker circles show parks with more programs, making it easy to spot busy parks versus quieter ones.

Where a park is located also matters. Neighborhoods that are larger or more central usually have more programs because they have more space, more facilities, and more visitors. With the filters on the left, anyone can explore the data by season, activity type, price, or park name. This makes the information easy to use even for someone with no data experience. For example, you can look for free programs, summer-only programs, or activities at a specific park in your neighborhood.

This project also highlights questions about access and opportunities. Some parks offer a wide range of programs, while others have fewer options or mostly offer only one type of activity. By looking at categories, seasons, and fees, readers can start to see patterns in which communities have more choices and which ones may need more support. Our goal is to turn public data into something simple and useful, so Chicago residents and decision-makers can better understand how parks are serving their communities.
""")

# -------------------------
# CITATIONS
# -------------------------
st.markdown("---")
st.subheader("Citations & Data Sources")
st.markdown("""
**Primary dataset:**  
Chicago Park District Activities — City of Chicago Data Portal  
https://data.cityofchicago.org/Parks-Recreation/Chicago-Park-District-Activities/tn7v-6rnw  
""")