""" Builds visualizations comparing the size of each D65 8th-grade cohort to the size of the same students' notional ETHS 9th-grade cohort the following year. The premise: every June, D65 graduates ~700-900 8th-graders. Every August, ETHS receives a freshman class. If parents trust D65 to deliver good K-8 education, most D65 8th-graders should continue on to ETHS. If they don't trust D65, they would be more likely to enroll children K-8 in private schools and bring them BACK to ETHS for high school — visible as ETHS Grade 9 being *larger* than the D65 8th-grade cohort the year before. Reads: data/d65_to_eths_cohorts.csv Writes: assets/k8_to_hs_*.html """ import sys import textwrap from pathlib import Path import pandas as pd import plotly.graph_objects as go from plotly.subplots import make_subplots ROOT = Path(__file__).parent DATA = ROOT / "data" ASSETS = ROOT / "assets" # write_html adds the phone layout (layout.meta.narrow); see tools/patch_plotly_layout.py. sys.path.insert(0, str(ROOT.parent / "tools")) from patch_plotly_layout import write_html as write_chart_html # noqa: E402 D65_COLOR = "#2980b9" # blue ETHS_COLOR = "#c0392b" # red — high contrast vs D65 GAIN_COLOR = "#e67e22" # orange — the "external gain" portion # ------------------------------------------------------------ # Load and derive # ------------------------------------------------------------ df = pd.read_csv(DATA / "d65_to_eths_cohorts.csv") # Year shown on charts: D65 8th-grade school year (the cohort) df["d65_end_year"] = df["d65_grade8_year"].str[-2:].astype(int) + 2000 df["eths_end_year"] = df["eths_grade9_year"].str[-2:].apply( lambda s: 2000 + int(s) if s.isdigit() else None) # Drop rows where ETHS data isn't yet available plot_df = df.dropna(subset=["eths_grade9_count"]).copy() plot_df["eths_grade9_count"] = plot_df["eths_grade9_count"].astype(int) plot_df["external_gain"] = plot_df["eths_grade9_count"] - plot_df["d65_grade8_count"] plot_df["gain_pct"] = plot_df["external_gain"] / plot_df["d65_grade8_count"] * 100 plot_df["ratio"] = plot_df["eths_grade9_count"] / plot_df["d65_grade8_count"] plot_df["cohort_label"] = plot_df["d65_grade8_year"] + " → " + plot_df["eths_grade9_year"] def wrap(text, width): return "
".join(textwrap.wrap(text, width)) def title_top(text): """Top margin for a bold 18 px title of this many lines.""" return 28 + text.count("
") * 24 + 34 def base_layout(title, xt="School Year (D65 8th-grade cohort, then ETHS 9th-grade)", yt="", height=550, legend_y=None): # The legend sits at the bottom of the figure (yref="container"), below the rotated year labels # and the axis title; legend_y is no longer used. text = f"{wrap(title, 64)}" return dict( title=dict(text=text, x=0.5, xanchor="center", font=dict(size=18), yref="container", y=1 - 28 / height, yanchor="top"), xaxis_title=xt, yaxis_title=yt, height=height, legend=dict(orientation="h", x=0.5, xanchor="center", yref="container", y=0, yanchor="bottom"), plot_bgcolor="white", xaxis=dict(showgrid=True, gridcolor="#eee", automargin=True), yaxis=dict(showgrid=True, gridcolor="#eee"), margin=dict(l=70, r=30, t=title_top(text), b=60), ) def write(fig, name, restyle=None): """Write the chart, with a phone layout: title wrapped narrower, x labels shortened to the D65 8th-grade year (the axis title says what the pair means), and bar labels hidden where they would have to shrink below 8 px.""" phone_title = f"{wrap(fig.layout.title.text[3:-4].replace('
', ' '), 28)}" cats = list(plot_df["cohort_label"]) narrow = { "title.text": phone_title, "margin.t": title_top(phone_title), "margin.l": 50, "margin.r": 10, "xaxis.tickvals": cats, "xaxis.ticktext": [c.split(" → ")[0] for c in cats], "xaxis.title.text": "D65 8th-grade year
(ETHS 9th grade the next fall)", "uniformtext.mode": "hide", "uniformtext.minsize": 8, "legend.font.size": 11, } if "yaxis2" in fig.layout.to_plotly_json(): # chart 2 has a right-hand axis narrow["margin.r"] = 50 out = ASSETS / f"k8_to_hs_{name}.html" write_chart_html(fig, out, narrow=narrow, restyle=restyle, include_plotlyjs="cdn", div_id=f"k8_to_hs_{name}") print(f"Wrote {out}") # ------------------------------------------------------------ # Chart 1 — Side-by-side cohort comparison # ------------------------------------------------------------ fig1 = go.Figure() fig1.add_trace(go.Bar( x=plot_df["cohort_label"], y=plot_df["d65_grade8_count"], name="D65 8th grade
(June graduating)", marker_color=D65_COLOR, text=plot_df["d65_grade8_count"], textposition="outside", hovertemplate="%{x}
D65 8th grade: %{y}", )) fig1.add_trace(go.Bar( x=plot_df["cohort_label"], y=plot_df["eths_grade9_count"], name="ETHS 9th grade
(following August)", marker_color=ETHS_COLOR, text=plot_df["eths_grade9_count"], textposition="outside", hovertemplate="%{x}
ETHS 9th grade: %{y}", )) fig1.update_layout(barmode="group", **base_layout( "D65 8th-grade Cohort vs. ETHS 9th-grade Cohort (one year later)", yt="Students enrolled (Sept 30 / Oct 1)")) fig1.update_layout(xaxis=dict(tickangle=-45)) write(fig1, "cohort_side_by_side") # ------------------------------------------------------------ # Chart 2 — External gain (ETHS 9th – D65 8th) over time # absolute counts, with a secondary y-axis for percentage # ------------------------------------------------------------ fig2 = make_subplots(specs=[[{"secondary_y": True}]]) fig2.add_trace(go.Bar( x=plot_df["cohort_label"], y=plot_df["external_gain"], name="Net external gain (students)", marker_color=GAIN_COLOR, text=[f"+{v}" for v in plot_df["external_gain"]], textposition="outside", hovertemplate="%{x}
External gain: +%{y} students", ), secondary_y=False) fig2.add_trace(go.Scatter( x=plot_df["cohort_label"], y=plot_df["gain_pct"], name="Gain as % of D65 8th-grade cohort", mode="lines+markers", line=dict(color="#34495e", width=3, dash="dot"), marker=dict(size=10, symbol="diamond"), hovertemplate="%{x}
Gain as %% of D65 cohort: %{y:.1f}%", ), secondary_y=True) fig2.update_yaxes(title_text="Net external gain (students)", secondary_y=False) fig2.update_yaxes(title_text="Gain as % of D65 cohort", ticksuffix="%", secondary_y=True) fig2.update_layout(**base_layout( "Net External Gain at the K-8 → High School Transition", yt="")) fig2.update_layout(xaxis=dict(tickangle=-45)) write(fig2, "external_gain_trend") # ------------------------------------------------------------ # Chart 3 — Stacked composition of ETHS 9th grade # "Came from D65 (estimate)" + "External gain" = ETHS 9th total # ------------------------------------------------------------ fig3 = go.Figure() fig3.add_trace(go.Bar( x=plot_df["cohort_label"], y=plot_df["d65_grade8_count"], name="Implied D65→ETHS continuers", marker_color=D65_COLOR, hovertemplate="%{x}
D65→ETHS continuers (assumed): %{y}", )) fig3.add_trace(go.Bar( x=plot_df["cohort_label"], y=plot_df["external_gain"], name="External gain (private K-8 returners + transfers)", marker_color=GAIN_COLOR, text=[f"+{v}" for v in plot_df["external_gain"]], textposition="inside", hovertemplate="%{x}
External gain: +%{y}", )) # Add total labels on top of each stack totals = plot_df["eths_grade9_count"].tolist() fig3.add_trace(go.Scatter( x=plot_df["cohort_label"], y=totals, mode="text", text=[str(v) for v in totals], textposition="top center", textfont=dict(size=12, color="#222"), showlegend=False, hoverinfo="skip", )) fig3.update_layout(barmode="stack", **base_layout( "Composition of Each ETHS Freshman Class", yt="ETHS 9th-grade enrollment")) fig3.update_layout(xaxis=dict(tickangle=-45)) # On phones the stack totals (trace 3, text only) collide; hide them there (the axis gives the total). write(fig3, "eths_composition", restyle={"mode": [None, None, "none"]}) # ------------------------------------------------------------ # Chart 4 — Trend in the gap (regression-style line chart) # ------------------------------------------------------------ fig4 = go.Figure() fig4.add_trace(go.Scatter( x=plot_df["cohort_label"], y=plot_df["gain_pct"], mode="lines+markers", line=dict(color=GAIN_COLOR, width=4), marker=dict(size=12, color=GAIN_COLOR), name="External gain as % of D65 cohort", hovertemplate="%{x}
%{y:.1f}%", )) # Mean line for reference mean_pct = plot_df["gain_pct"].mean() fig4.add_hline(y=mean_pct, line=dict(color="#999", width=1, dash="dash"), annotation_text=f"10-yr mean: {mean_pct:.1f}%", annotation_position="top left") fig4.update_layout(**base_layout( "Has the K-8 Trust Gap Grown? — External Gain as % of D65 Cohort, by Year", yt="External gain as % of D65 cohort", height=530)) # 530 + page margins fits the page's 550 px frame fig4.update_yaxes(ticksuffix="%") fig4.update_layout(xaxis=dict(tickangle=-45)) write(fig4, "trend_pct") # ------------------------------------------------------------ # Print summary (ASCII-only labels for Windows console safety) # ------------------------------------------------------------ ascii_summary = plot_df.copy() ascii_summary["cohort_label"] = ascii_summary["cohort_label"].str.replace("→", "->", regex=False) print("\n=== Cohort summary ===") print(ascii_summary[["cohort_label", "d65_grade8_count", "eths_grade9_count", "external_gain", "gain_pct"]].to_string(index=False)) print(f"\n10-year mean external gain: {plot_df['external_gain'].mean():.0f} students " f"({mean_pct:.1f}% of D65 cohort)") min_label = ascii_summary.loc[plot_df['external_gain'].idxmin(), 'cohort_label'] max_label = ascii_summary.loc[plot_df['external_gain'].idxmax(), 'cohort_label'] print(f"Min: {plot_df['external_gain'].min()} ({plot_df['gain_pct'].min():.1f}%) in {min_label}") print(f"Max: {plot_df['external_gain'].max()} ({plot_df['gain_pct'].max():.1f}%) in {max_label}") print("Done.")