Skip to content

Latest commit

 

History

History
72 lines (36 loc) · 2.71 KB

README.md

File metadata and controls

72 lines (36 loc) · 2.71 KB
Raw

Threshold Simulator

User Story

Who? Persona Strengths

Cash-Based-Intervention Focal points in Operations

What? Goal: what they need to do?

Make an informed decision on what weights and scoring threshold identifies the most vulnerable profiles within budget constraints.

Why? Context: In which kind of situation?

Visualize how different household profiles affect vulnerability scores, which then impact eligibility

Frustration: what prevents doing their work​

Currently... Lack of methodology/tool to visualize impact of different thresholds (and weights) in the beneficiaries list.

Motivation: benefit – expected user gain –

Ideally...The tool will show scores for the three components, and a final composite score. With option to modify cutoff threshold to see the impact in number of eligible profiles. ​ Ability to modify the weights of the three components in the final score and see the impact in eligible profiles. ​

Ability to visualize scores based on specific profiles : age, gender, household size, specific needs and coping mechanisms, geographic zone.

Specification and technical requirement

Stage 1: Develop a Static Notebook with parameters using key functions

Stage2: validate with users the static version

Stage 3: Develop a Dynamic dashboard with ability for users to change parameters

1. Input data:

Upload Assessment dataset on the App...

2. User input: (dynamic control from CBI officer...)

  • Confirm the variable name for the 3 dimensions

  • Select key variables for visualizations (choose multiple from existing variable in data)

  • Set up and Select as per usage scenario:

    • Scenario 1: Sliders with weights for each dimensions to aggregate in one unique score – then threshold

    • Scenario 2: Sliders with eligibility and prioritization thresholds for each of the 3 dimensions

3. Output viz to facilitate iteration based on user input:

  • Display key variable to display as faceted charts – prioritized vs not_prioritized

  • Key metrics – Total # prioritized - # eligible

4. Data extract at the end of the research iteration:

  • All original data with additional column - "Not eligible", "Eligible", "Prioritized" (plus composite unique score in case of usage scenario 1)​

  • Additional worksheet with metadata on user input decision

Main Functions

`write_prioritisation_decision()` -> this will write the user input (metadata) in a selected object​

`calculate_prioritisation()` -> create a new `prioritisation` variable based on the metadata​

`show_profiles()` -> ggplot2 facet chart based on the list of defined variable to visualise faceted based on `prioritisation` variable