Pre-baked scenarios

README.md

@@ -16,7 +16,7 @@ approach.
 
 ## Model Description
 
-This repo contains code to apply the next-generation method of Diekman et al. (1990) to calculate R0 for an SIR model with 4 risk groups and flexible inputs for varying vaccine allocation to the 4 groups.
+This repo contains code to apply the next-generation method of Diekman et al. (1990) to calculate R0 for an SIR model with 3 risk groups and flexible inputs for varying vaccine allocation to each group.
 
 ### Next generation matrix calculation
 
@@ -33,13 +33,12 @@ Default parameters in the widget assume the following about transmission:
 | --------- | ------------------------------------------------ | ------------------------------------------------ | -------------------------------------- |
 | Core      | High                                             | High                                             | Low                                    |
 | Children  | Low                                              | Low                                              | High                                   |
-| Travelers | Low                                              | High                                             | Low                                    |
 | General   | Low                                              | Low                                              | Low                                    |
 
 Other Inputs to the widget:
 
 - Sizes of the groups, $N_i$
-- Vaccination efficacy (VE) and number of doses allocated to each group ($V_i$)
+- Vaccination efficacy (VE) and number of doses allocated to each group ($v_i$)
 - Within and between group contact rates; $\beta_{ij}$
 - Per-group probability of severe infection
 

docs/approach.md

@@ -1,8 +1,8 @@
 # Using Next Generation Matrices
 
-The NGM for a 4-Group Infectious Disease Model with compartments $S_i$, $I_i$, $R_i$: Susceptible, Infected, and Recovered compartments.
+The NGM for a 3-Group Infectious Disease Model with compartments $S_i$, $I_i$, $R_i$: Susceptible, Infected, and Recovered compartments.
 
-The dynamics for $i$ in each group given by:
+The dynamics for $I$ in each group given by:
 
 $$
 \frac{d I_i}{dt} = \sum_{j} \frac{\beta_{ij} S_i I_j}{N} - \gamma_i I_i
@@ -21,31 +21,31 @@ $$
 I_i = 0, S_i \approx N_i \  \text{for all\ } i
 $$
 
-And then the NGM $\mathbf{R}_{ij}$ is given by:
+The NGM $\mathbf{R}_{ij}$ is given by:
 
 $$
 \mathbf{R}_{ij} = \frac{\beta_{ij} S_{i}}{\gamma N}
 $$
 
-If all $\gamma_i = \gamma$ and we assume SIR dynamics.
+Where we assume all $\gamma_i = \gamma$ and SIR dynamics.
 
-The basic reproductive number $R_0$ is calculated as the dominant eigenvalue of $R$.
+The basic reproductive number $R_0$ is calculated as the [spectral radius](https://en.wikipedia.org/wiki/Spectral_radius) (the largest absolute value of the eigenvalues) of $\mathbf{R}$.
 
-This model incorporates vaccination by recalculating the distribution of susceptible individuals in each group $S_{i}^{vax}$ (assuming all or nothing vaccination, with vaccine efficacy given by $ve$ and the proportion of $i$ vacinated is $v_i$):
+This model incorporates vaccination by recalculating the number of susceptible individuals in each group $S_{i}^{\mathrm{vax}}$ (assuming all or nothing vaccination, with vaccine efficacy given by $\mathrm{VE}$ and the proportion of $i$ vacinated is $v_i$):
 
 $$
 \mathbf{S_{i}^\mathrm{vax}} = S_{i} - v_{i} * \mathrm{VE}
 $$
 
-So that $S_i^\mathrm{vax}$ is the population $i$ that is still susceptible post vaccination administration. Then \mathbf{R}_{ij} with vaccination factored in is given by
+So that $S_i^\mathrm{vax}$ is the population $i$ that is still susceptible post vaccination administration. Then $\mathbf{R}_{ij}$ is, accounting for vaccination,
 
 $$
-\mathbf{R}_{ij}^{vax} = \mathbf{R}_{ij} \frac{S_{i}^\mathrm{vax}}{N_i}
+\mathbf{R}_{ij}^\mathrm{vax} = \mathbf{R}_{ij} \frac{S_{i}^\mathrm{vax}}{N_i}
 $$
 
 
-The dominant eigenvalue of $R$ with vaccination is $R_e$.
+The spectral radius of $\mathbf{R}^\mathrm{vax}$ is $R_e$.
 
-The distribution of infections is calculated from the dominant eigenvector of $R$ (specifying that its elements add to 1).
+The distribution of infections is calculated from the eigenvector that corresponds to the spectral radius of $\mathbf{R}^\mathrm{vax}$ (specifying that the elements of said eigenvector sum to 1).
 
 Severe infections are calculated by multiplying the proportion of infections in each group by a group-specific probability of severe infection.

widget/widget.py

@@ -7,17 +7,17 @@
 
 
 def app():
-    st.title("4-Group NGM Calculator")
+    st.title("3-Group NGM Calculator")
 
     # Group names
-    group_names = ["Core", "Kids", "Travelers", "General Population"]
+    group_names = ["Core", "Kids", "General Population"]
 
     # Sidebar for inputs
     st.sidebar.header("Model Inputs")
 
     # Population size
     st.sidebar.subheader("Population Sizes")
-    default_values = np.array([0.2, 0.2, 0.005, 0.595]) * 1000000
+    default_values = np.array([0.2, 0.2, 0.595]) * 1000000
     N = np.array(
         [
         st.sidebar.number_input(f"Population ({group})", value=int(default_values[i]), min_value=0)
@@ -26,12 +26,26 @@ def app():
     )
 
     # Vaccine doses
-    st.sidebar.subheader("Vaccine Doses")
+
+    ndoses = 100000
+    st.sidebar.subheader("Vaccine Doses (Pre-filled Scenarios)")
+    starting_vax = st.sidebar.selectbox("Vaccine allocation strategy", ["No vaccination", "All core", "All kids", "Even"])
+    # No vax is default
+    allocation = [0, 0, 0]
+    if starting_vax == "All core":
+        allocation = [1, 0, 0]
+    elif starting_vax == "All kids":
+        allocation = [0, 1, 0]
+    elif starting_vax == "Even":
+        allocation = [1/3, 1/3, 1/3]
+    allocation = np.floor(np.array(allocation) * ndoses).astype("int")
+
+    st.sidebar.subheader("Vaccine Doses (Customization)")
     V = np.array(
         [
             st.sidebar.number_input(
                 f"Vaccine Doses ({group})",
-                value=0,
+                value=allocation[i],
                 min_value=0,
                 max_value=N[i],
             )
@@ -48,10 +62,9 @@ def app():
     # Create the contact matrix K
     beta = np.array(
         [
-            [hi, lo, hi, lo],  # core
-            [lo, lo, lo, lo],  # kids
-            [hi, lo, lo, lo],  # travelers
-            [lo, lo, lo, lo],  # general
+            [hi, lo, lo],  # core
+            [lo, lo, lo],  # kids
+            [lo, lo, lo],  # general
         ]
     )
 
@@ -61,7 +74,7 @@ def app():
 
     # Perform the NGM calculation
     result = ngm.simulate(
-        n=N, n_vax=V, beta=beta, p_severe=np.array([0.02, 0.06, 0.02, 0.02]), ve=VE
+        n=N, n_vax=V, beta=beta, p_severe=np.array([0.02, 0.06, 0.02]), ve=VE
     )
 
     # Display the adjusted contact matrix