Update app.py

app.py

@@ -5,14 +5,13 @@ import yfinance as yf
 import datetime
 import plotly.graph_objs as go
 
-# ---------- Page config (must be the first Streamlit call) ----------
+# ---------- Page config (must be first) ----------
 st.set_page_config(layout="wide")
 
-# ---------- Stable CSS for wider sidebar (avoid fragile class names) ----------
+# ---------- Stable CSS for wider sidebar ----------
 st.markdown(
     """
     <style>
-    /* Make the sidebar wider in a stable way */
     [data-testid="stSidebar"] {
         width: 350px;
         min-width: 350px;
@@ -22,11 +21,15 @@ st.markdown(
     unsafe_allow_html=True,
 )
 
-# ---------- Session state for persistent "Run Analysis" ----------
+# ---------- Session state ----------
 if "run_analysis" not in st.session_state:
     st.session_state.run_analysis = False
+if "combined_data" not in st.session_state:
+    st.session_state.combined_data = None
+if "selection_signature" not in st.session_state:
+    st.session_state.selection_signature = None
 
-# ---------- App title and description (kept as-is) ----------
+# ---------- App title / description ----------
 st.title("Key Economic Recession Indicators")
 st.markdown("""
 This tool allows you to visualize and analyze various recession indicators over time.
@@ -34,7 +37,7 @@ This tool allows you to visualize and analyze various recession indicators over
 - Use the checkboxes in the sidebar to choose the indicators you'd like to explore.
 """)
 
-# ---------- Sidebar controls ----------
+# ---------- Sidebar ----------
 with st.sidebar.expander("How to Use", expanded=False):
     st.write("""
     **How to use this app:**
@@ -50,8 +53,8 @@ with st.sidebar.expander("Indicators", expanded=True):
         'Sahm Recession Indicator': 'SAHMREALTIME',
         'U.S. Recession Probabilities': 'RECPROUSM156N',
         'Yield Spread (10Y - 2Y)': 'Yield_Spread',  # Calculated, not fetched
-        'Stock Market (S&P 500)': 'SP500',          # Fetched from yfinance
-        'VIX': 'VIX',                               # Fetched from yfinance
+        'Stock Market (S&P 500)': 'SP500',          # yfinance (^GSPC)
+        'VIX': 'VIX',                               # yfinance (^VIX)
         'Treasury Rates': ('GS10', 'DGS2', 'DGS1MO', 'TB3MS'),
         'Federal Funds Rate': 'FEDFUNDS',
         'Unemployment Rate': 'UNRATE',
@@ -65,7 +68,7 @@ with st.sidebar.expander("Indicators", expanded=True):
     }
     selected_indicators = {key: st.checkbox(key, value=True) for key in indicators.keys()}
 
-# Single Run button (no clear buttons)
+# Single Run button
 if st.sidebar.button("Run Analysis"):
     st.session_state.run_analysis = True
 
@@ -73,7 +76,7 @@ if st.sidebar.button("Run Analysis"):
 start_date = datetime.datetime(1920, 1, 1)
 end_date = datetime.datetime.today()
 
-# ---------- Recession periods (kept as-is) ----------
+# ---------- Recession periods ----------
 crash_periods = {
     '1929-08-01': '1933-03-01',
     '1937-05-01': '1938-06-01',
@@ -92,154 +95,192 @@ crash_periods = {
     '2020-02-01': '2020-04-01'
 }
 
-# ---------- Cached data fetchers ----------
-@st.cache_data(ttl=6 * 60 * 60, show_spinner=False)
-def fetch_fred_series(series_code: str, start: datetime.datetime, end: datetime.datetime) -> pd.Series:
-    """Fetch a single FRED series as a named Series (empty Series if fails)."""
+# ---------- Cached batch fetchers ----------
+@st.cache_data(ttl=6*60*60, show_spinner=False)
+def fetch_fred_batch(series_codes, start, end) -> pd.DataFrame:
+    """Batch fetch from FRED; falls back to per-series if batch fails."""
+    if not series_codes:
+        return pd.DataFrame()
     try:
-        df = web.DataReader(series_code, 'fred', start, end)
-        if isinstance(df, pd.DataFrame):
-            s = df.squeeze("columns")
-        else:
-            s = df
-        s = s.rename(series_code)
-        return s
+        df = web.DataReader(series_codes, "fred", start, end)
+        # DataReader may return a Series when series_codes has length 1
+        if isinstance(df, pd.Series):
+            df = df.to_frame(series_codes[0])
+        return df
     except Exception as e:
-        st.warning(f"Failed to fetch {series_code} from FRED: {e}")
-        return pd.Series(name=series_code, dtype="float64")
-
-@st.cache_data(ttl=6 * 60 * 60, show_spinner=False)
-def fetch_yf_series(ticker: str, label: str, start: datetime.datetime, end: datetime.datetime) -> pd.Series:
-    """Fetch Adj Close from Yahoo Finance as a named Series."""
+        # Fallback: try individually, salvage what we can
+        frames = []
+        for code in series_codes:
+            try:
+                s = web.DataReader(code, "fred", start, end).squeeze("columns").rename(code)
+                frames.append(s)
+            except Exception as ie:
+                st.warning(f"Failed to fetch {code} from FRED: {ie}")
+        return pd.concat(frames, axis=1) if frames else pd.DataFrame()
+
+@st.cache_data(ttl=6*60*60, show_spinner=False)
+def fetch_yf_batch(tickers, start, end) -> pd.DataFrame:
+    """Batch fetch Adj Close from Yahoo Finance, rename to friendly labels."""
+    if not tickers:
+        return pd.DataFrame()
     try:
-        df = yf.download(ticker, start=start, end=end, auto_adjust=False, progress=False, threads=False)
+        df = yf.download(
+            tickers=" ".join(tickers),
+            start=start, end=end,
+            auto_adjust=False, progress=False, threads=False,
+        )
+        # Normalize to a flat DataFrame of Adj Close columns
         if isinstance(df.columns, pd.MultiIndex):
-            df.columns = df.columns.get_level_values(0)
-        s = df.get('Adj Close', pd.Series(dtype="float64")).rename(label)
-        return s
+            # pick Adj Close for each ticker
+            if "Adj Close" in df.columns.get_level_values(0):
+                wide = df["Adj Close"].copy()
+            else:
+                # Fallback: try Close
+                wide = df["Close"].copy()
+        else:
+            # Single ticker -> just one column set
+            wide = df[["Adj Close"]] if "Adj Close" in df.columns else df[["Close"]]
+            wide.columns = [tickers[0]]
+        # Rename to app labels
+        rename_map = {"^GSPC": "SP500", "^VIX": "VIX"}
+        wide = wide.rename(columns=rename_map)
+        return wide
     except Exception as e:
-        st.warning(f"Failed to fetch {label} ({ticker}) from Yahoo Finance: {e}")
-        return pd.Series(name=label, dtype="float64")
+        st.warning(f"Failed to fetch from Yahoo Finance: {e}")
+        return pd.DataFrame()
+
+# ---------- Build everything up-front ----------
+def selection_signature(selected: dict) -> tuple:
+    """Immutable key for the current selection (used to decide if we refetch)."""
+    return tuple(sorted([k for k, v in selected.items() if v]))
 
-# ---------- Build dataset ----------
-def build_dataset(selected: dict) -> pd.DataFrame:
-    series_list = []
+@st.cache_data(ttl=6*60*60, show_spinner=False)
+def build_dataset_all(selected_keys: tuple) -> pd.DataFrame:
+    """Build the full dataset for the given selection (batched)."""
+    # Recreate indicator mapping (cache functions need pure inputs)
+    indicators_local = {
+        'Sahm Recession Indicator': 'SAHMREALTIME',
+        'U.S. Recession Probabilities': 'RECPROUSM156N',
+        'Yield Spread (10Y - 2Y)': 'Yield_Spread',
+        'Stock Market (S&P 500)': 'SP500',
+        'VIX': 'VIX',
+        'Treasury Rates': ('GS10', 'DGS2', 'DGS1MO', 'TB3MS'),
+        'Federal Funds Rate': 'FEDFUNDS',
+        'Unemployment Rate': 'UNRATE',
+        'Nonfarm Payrolls': 'PAYEMS',
+        'Jobless Claims': 'ICSA',
+        'Retail Sales': 'RSXFS',
+        'Industrial Production': ('INDPRO', 'INDPRO_PCT'),
+        'Housing Starts': 'HOUST',
+        'Consumer Confidence': 'UMCSENT',
+        'Inflation (CPI)': ('CPIAUCSL', 'CPIAUCSL_PCT')
+    }
 
-    # FRED (skip derived)
-    for key, col in indicators.items():
-        if not selected.get(key, False):
-            continue
+    fred_series = []
+    yf_tickers = []
+    for key in selected_keys:
+        col = indicators_local[key]
         if isinstance(col, tuple):
             for c in col:
-                if c in ["INDPRO_PCT", "CPIAUCSL_PCT"]:
-                    continue  # derived later
-                s = fetch_fred_series(c, start_date, end_date)
-                if not s.empty:
-                    series_list.append(s)
+                if c in ("INDPRO_PCT", "CPIAUCSL_PCT"):  # derived later
+                    continue
+                fred_series.append(c)
         else:
-            if col in ["Yield_Spread", "SP500", "VIX"]:
-                continue  # handled separately / derived
-            s = fetch_fred_series(col, start_date, end_date)
-            if not s.empty:
-                series_list.append(s)
-
-    # YFinance
-    if selected.get('Stock Market (S&P 500)', False):
-        s = fetch_yf_series('^GSPC', 'SP500', start_date, end_date)
-        if not s.empty:
-            series_list.append(s)
-    if selected.get('VIX', False):
-        s = fetch_yf_series('^VIX', 'VIX', start_date, end_date)
-        if not s.empty:
-            series_list.append(s)
-
-    if not series_list:
+            if col == "SP500":
+                yf_tickers.append("^GSPC")
+            elif col == "VIX":
+                yf_tickers.append("^VIX")
+            elif col == "Yield_Spread":
+                pass  # derived later
+            else:
+                fred_series.append(col)
+
+    fred = fetch_fred_batch(sorted(set(fred_series)), start_date, end_date)
+    yfdf = fetch_yf_batch(sorted(set(yf_tickers)), start_date, end_date)
+
+    frames = []
+    if isinstance(fred, pd.DataFrame) and not fred.empty:
+        frames.append(fred)
+    if isinstance(yfdf, pd.DataFrame) and not yfdf.empty:
+        frames.append(yfdf)
+
+    if not frames:
         return pd.DataFrame()
 
-    combined = pd.concat(series_list, axis=1).sort_index()
+    combined = pd.concat(frames, axis=1).sort_index()
 
-    # Derived columns
-    if selected.get('Industrial Production', False) and 'INDPRO' in combined.columns:
-        combined['INDPRO_PCT'] = combined['INDPRO'].pct_change() * 100
-    if selected.get('Inflation (CPI)', False) and 'CPIAUCSL' in combined.columns:
-        combined['CPIAUCSL_PCT'] = combined['CPIAUCSL'].pct_change() * 100
+    # Derived series
+    if "INDPRO" in combined.columns and "Industrial Production" in selected_keys:
+        combined["INDPRO_PCT"] = combined["INDPRO"].pct_change() * 100
+    if "CPIAUCSL" in combined.columns and "Inflation (CPI)" in selected_keys:
+        combined["CPIAUCSL_PCT"] = combined["CPIAUCSL"].pct_change() * 100
 
-    # Interpolate (time index required)
-    combined = combined.interpolate(method='time')
+    # Interpolate for alignment
+    combined = combined.interpolate(method="time")
 
     # Yield spread
-    if selected.get('Yield Spread (10Y - 2Y)', False) and {'GS10', 'DGS2'}.issubset(combined.columns):
-        combined['Yield_Spread'] = combined['GS10'] - combined['DGS2']
+    if "Yield Spread (10Y - 2Y)" in selected_keys and {"GS10", "DGS2"}.issubset(combined.columns):
+        combined["Yield_Spread"] = combined["GS10"] - combined["DGS2"]
 
     return combined
 
-# ---------- Plotting helpers ----------
+# ---------- Plot helpers ----------
 def add_recession_shading(fig: go.Figure):
     for peak, trough in crash_periods.items():
         fig.add_shape(
-            type="rect",
-            xref="x",
-            yref="paper",
-            x0=peak,
-            y0=0,
-            x1=trough,
-            y1=1,
-            fillcolor="gray",
-            opacity=0.3,
-            layer="below",
-            line_width=0,
+            type="rect", xref="x", yref="paper",
+            x0=peak, y0=0, x1=trough, y1=1,
+            fillcolor="gray", opacity=0.3, layer="below", line_width=0,
         )
 
 def finalize_layout(fig: go.Figure, title: str, ytitle: str):
     fig.update_layout(
         title=title,
-        xaxis_title='Date',
+        xaxis_title="Date",
         yaxis_title=ytitle,
         xaxis=dict(
-            tickformat="%Y",
-            tickmode="linear",
-            dtick="M36",
-            showspikes=True,
-            spikemode='across',
-            spikesnap='cursor',
-            spikethickness=1
+            tickformat="%Y", tickmode="linear", dtick="M36",
+            showspikes=True, spikemode="across", spikesnap="cursor", spikethickness=1
         ),
         hovermode="x unified",
-        hoverlabel=dict(
-            bgcolor="white",
-            font_size=12,
-            font_family="Rockwell"
-        ),
+        hoverlabel=dict(bgcolor="white", font_size=12, font_family="Rockwell"),
         legend=dict(
-            x=0.02,
-            y=0.95,
-            traceorder='normal',
-            bgcolor='rgba(255, 255, 255, 0.5)',
-            bordercolor='rgba(0, 0, 0, 0)'
+            x=0.02, y=0.95, traceorder="normal",
+            bgcolor="rgba(255,255,255,0.5)", bordercolor="rgba(0,0,0,0)"
         ),
-        margin=dict(l=60, r=20, t=60, b=40)
+        margin=dict(l=60, r=20, t=40, b=40),
     )
     fig.update_xaxes(
-        showgrid=True,
-        gridwidth=1,
-        gridcolor='LightGray',
-        tickangle=45,
+        showgrid=True, gridwidth=1, gridcolor="LightGray", tickangle=45,
         tickformatstops=[
             dict(dtickrange=[None, "M1"], value="%b %d, %Y"),
-            dict(dtickrange=["M1", None], value="%Y")
-        ]
+            dict(dtickrange=["M1", None], value="%Y"),
+        ],
     )
-    fig.update_traces(hovertemplate='%{x|%b %d, %Y}<br>%{y}<extra></extra>')
+    fig.update_traces(hovertemplate="%{x|%b %d, %Y}<br>%{y}<extra></extra>")
 
 # ---------- Main render ----------
 if st.session_state.run_analysis:
-    with st.spinner("Fetching data and building charts..."):
-        combined_data = build_dataset(selected_indicators)
+    # Compute a signature of the current selection to decide if we need to rebuild
+    sig = selection_signature(selected_indicators)
 
-    if combined_data.empty:
+    need_fetch = (
+        st.session_state.combined_data is None
+        or st.session_state.selection_signature != sig
+    )
+
+    if need_fetch:
+        with st.spinner("Fetching all selected data (batched) and building dataset..."):
+            combined_data = build_dataset_all(sig)
+        st.session_state.combined_data = combined_data
+        st.session_state.selection_signature = sig
+    else:
+        combined_data = st.session_state.combined_data
+
+    if combined_data is None or combined_data.empty:
         st.error("No data was successfully fetched for the selected indicators.")
     else:
-        # Loop through selections and plot
+        # Plot everything AFTER data is fully ready
         for key, column in indicators.items():
             if not selected_indicators.get(key, False):
                 continue
@@ -248,7 +289,7 @@ if st.session_state.run_analysis:
             add_recession_shading(fig)
 
             if isinstance(column, tuple):
-                # Industrial Production: level + % change on y2
+                # Industrial Production (level + % change on y2)
                 if column == ('INDPRO', 'INDPRO_PCT') and 'INDPRO' in combined_data.columns:
                     fig.add_trace(go.Scatter(
                         x=combined_data.index, y=combined_data['INDPRO'],
@@ -259,11 +300,12 @@ if st.session_state.run_analysis:
                             x=combined_data.index, y=combined_data['INDPRO_PCT'],
                             mode='lines', name='Industrial Production % Change', yaxis='y2'
                         ))
-                        fig.update_layout(yaxis2=dict(title="Industrial Production % Change",
-                                                      overlaying='y', side='right'))
+                        fig.update_layout(yaxis2=dict(
+                            title="Industrial Production % Change", overlaying='y', side='right'
+                        ))
                     finalize_layout(fig, key, key)
 
-                # Inflation: CPI + % change on y2
+                # Inflation CPI (level + % change on y2)
                 elif column == ('CPIAUCSL', 'CPIAUCSL_PCT') and 'CPIAUCSL' in combined_data.columns:
                     fig.add_trace(go.Scatter(
                         x=combined_data.index, y=combined_data['CPIAUCSL'],
@@ -274,24 +316,26 @@ if st.session_state.run_analysis:
                             x=combined_data.index, y=combined_data['CPIAUCSL_PCT'],
                             mode='lines', name='Inflation % Change', yaxis='y2'
                         ))
-                        fig.update_layout(yaxis2=dict(title="Inflation % Change",
-                                                      overlaying='y', side='right'))
+                        fig.update_layout(yaxis2=dict(
+                            title="Inflation % Change", overlaying='y', side='right'
+                        ))
                     finalize_layout(fig, key, key)
 
-                # Treasury rates: plot each available
+                # Treasury rates
                 elif column == ('GS10', 'DGS2', 'DGS1MO', 'TB3MS'):
                     any_added = False
                     for col in column:
                         if col in combined_data.columns:
-                            any_added = True
                             fig.add_trace(go.Scatter(
-                                x=combined_data.index, y=combined_data[col], mode='lines', name=col
+                                x=combined_data.index, y=combined_data[col],
+                                mode='lines', name=col
                             ))
+                            any_added = True
                     if any_added:
                         finalize_layout(fig, key, key)
 
                 else:
-                    # Generic multi-series if needed
+                    # Generic multi-series, if ever needed
                     for col in column:
                         if col in combined_data.columns:
                             fig.add_trace(go.Scatter(
@@ -320,15 +364,16 @@ if st.session_state.run_analysis:
                     ))
                     finalize_layout(fig, key, key)
 
-            # Only render if we actually added something beyond the shading
             if fig.data:
                 st.plotly_chart(fig, use_container_width=True)
 
-# ---------- Hide default Streamlit branding ----------
-hide_streamlit_style = """
-<style>
-#MainMenu {visibility: hidden;}
-footer {visibility: hidden;}
-</style>
-"""
-st.markdown(hide_streamlit_style, unsafe_allow_html=True)
+# ---------- Hide Streamlit branding ----------
+st.markdown(
+    """
+    <style>
+    #MainMenu {visibility: hidden;}
+    footer {visibility: hidden;}
+    </style>
+    """,
+    unsafe_allow_html=True,
+)