hivsim_eswatini/plot_fig2_hiv_by_age.py at main · starsimhub/hivsim_eswatini · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
"""
HIV prevalence by age and sex, compared to PHIA survey data.

Four stacked panels, one per survey year (2007, 2011, 2016, 2021).
Bars = model prevalence, diamonds = PHIA data with 95% CIs.
"""

import sciris as sc
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec
from utils import set_font

RESULTS_DIR = 'results'
FIGURES_DIR = 'figures'

# Colors
F_COLOR = '#d46e9c'
M_COLOR = '#4a90d9'

# Sim age bins and the PHIA 5-year bins they map to
SIM_BINS = [
    (15, 20, [15]),
    (20, 25, [20]),
    (25, 30, [25]),
    (30, 35, [30]),
    (35, 50, [35, 40, 45]),
    (50, 65, [50, 55, 60]),
]


def aggregate_phia(phia, year, gender, sim_lo, sim_hi, phia_starts):
    """Average PHIA 5-year bins into a sim bin, weighted by Count."""
    rows = phia[(phia.Year == year) & (phia.Gender == gender) &
                (phia['start age'].isin(phia_starts))]
    if len(rows) == 0:
        return np.nan, np.nan, np.nan

    counts = rows['Count'].values
    prev = rows['NationalPrevalence'].values
    lb = rows['lb'].values
    ub = rows['ub'].values

    if np.any(np.isnan(counts)) or np.sum(counts) == 0:
        # Simple average if counts missing
        return np.nanmean(prev), np.nanmean(lb), np.nanmean(ub)

    w = counts / counts.sum()
    return np.average(prev, weights=w), np.average(lb, weights=w), np.average(ub, weights=w)


def plot_hiv_by_age():
    sim_df = sc.loadobj(f'{RESULTS_DIR}/eswatini_sim.df')
    phia = pd.read_csv('raw_data/SWAZILAND_nationalprevalence_all_updatedPHIA3.csv')

    survey_years = [2007, 2011, 2016, 2021]
    bin_labels = [f'{lo}\u2013{hi}' for lo, hi, _ in SIM_BINS]
    x = np.arange(len(SIM_BINS))
    width = 0.35

    set_font(size=18)
    fig = plt.figure(figsize=(10, 16))
    gs = GridSpec(len(survey_years), 1, left=0.10, right=0.95, bottom=0.05, top=0.95, hspace=0.25)

    ymax = 70

    for idx, year in enumerate(survey_years):
        ax = fig.add_subplot(gs[idx])

        # Find closest sim year
        sim_years = sim_df['timevec'].values
        sim_year = sim_years[np.argmin(np.abs(sim_years - year))]

        sim_row = sim_df[sim_df['timevec'] == sim_year].iloc[0]

        # Model bars
        f_vals, m_vals = [], []
        for lo, hi, _ in SIM_BINS:
            f_col = f'hiv.prevalence_f_{lo}_{hi}'
            m_col = f'hiv.prevalence_m_{lo}_{hi}'
            f_vals.append(sim_row.get(f_col, 0) * 100)
            m_vals.append(sim_row.get(m_col, 0) * 100)

        f_label = 'Female' if idx == 0 else None
        m_label = 'Male' if idx == 0 else None
        ax.bar(x - width/2, f_vals, width, color=F_COLOR, alpha=0.85, label=f_label)
        ax.bar(x + width/2, m_vals, width, color=M_COLOR, alpha=0.85, label=m_label)

        # PHIA data overlay
        kw = dict(marker='D', s=50, zorder=5, edgecolors='k', linewidths=0.5)
        first = True
        for j, (lo, hi, phia_starts) in enumerate(SIM_BINS):
            for gender, sex_color, offset in [(1, F_COLOR, -width/2), (0, M_COLOR, width/2)]:
                val, lb, ub = aggregate_phia(phia, year, gender, lo, hi, phia_starts)
                if np.isnan(val):
                    continue
                label = f'PHIA {year}' if first else None
                ax.scatter(x[j] + offset, val * 100, color=sex_color, label=label, **kw)
                ax.errorbar(x[j] + offset, val * 100,
                            yerr=[[(val - lb) * 100], [(ub - val) * 100]],
                            fmt='none', ecolor=sex_color, capsize=2, linewidth=1, alpha=0.7)
                first = False

        ax.set_ylim(0, ymax)
        ax.text(0.98, 0.88, str(year), transform=ax.transAxes, ha='right', va='top',
                fontsize=16, fontweight='bold')

        if idx == len(survey_years) - 1:
            ax.set_xticks(x)
            ax.set_xticklabels(bin_labels)
            ax.set_xlabel('Age group')
        else:
            ax.set_xticks(x)
            ax.set_xticklabels([])

        if idx == 0:
            ax.set_title('HIV prevalence by age and sex')
            ax.legend(frameon=False, fontsize=14, loc='upper left', ncol=3)

        if idx == len(survey_years) // 2:
            ax.set_ylabel('Prevalence (%)')

    plt.savefig(f'{FIGURES_DIR}/fig2_hiv_by_age.png', dpi=200, bbox_inches='tight')
    print(f'Saved {FIGURES_DIR}/fig2_hiv_by_age.png')


if __name__ == '__main__':
    plot_hiv_by_age()