AI Sentiment Analysis Consultant

Population Reference Bureau

AI Sentiment Analysis Consultant (Temporary Position)

Initial Duration: January 10, 2026 - March 25, 2026

Number of Days: 20 days

About PRB:

PRB is a nonpartisan, not-for-profit research organization focused on improving people's health and well-being through evidence-based policies and practices. Our staff analyze population data and information and ensure the research, and its applications are understood and used widely by decisionmakers, advocates, and media.

Scope of work

The demographic transition (DT) has significant implications for economic development, gender equality, and sexual and reproductive health and rights (SRHR). Public sentiment around DT, SRHR, and contraception shapes the political and policy environment in which governments make investment and programming decisions. However, public discourse on these topics is often fragmented, influenced by political rhetoric, cultural norms, misinformation, and evolving narratives in both traditional and social media. To inform advocacy strategies that generate demand for government investment in SRHR and contraception, there is a need to systematically understand how the public perceives and discusses these issues, and how narratives evolve in response to major events.

Objective

The WISH project will commission a Uganda-based consultant to design and pilot an AI-driven public sentiment and narrative analysis to fill this gap. The analysis will:

1. Determine whether and how Ugandans discuss demographic transition (DT), SRHR, and contraception in traditional and social media (English-language content only)

2. Identify dominant framing patterns (e.g., women's empowerment, health, religious/moral concerns, economic arguments)

3. Test whether discourse shifts in response to policy events (e.g., release of a new family planning policy)

4. Generate actionable recommendations for advocacy messaging and policy engagement

Primary Responsibilities

The consultant will:

1. Identify and sample data sources

o Identify relevant English-language traditional media (newspapers, radio transcripts, TV) and social media sources (Twitter/X, Facebook public pages, online forums)

o Develop a sampling strategy that captures discourse before and after a defined policy event

o Document inclusion/exclusion criteria and any limitations

2. Collect, clean, and prepare data

o Acquire sample dataset (size to be agreed based on source availability)

o Standardize format for analysis

o Document preprocessing steps

3. Design sentiment and framing classification framework

o Develop classification categories beyond simple positive/negative/neutral - e.g., framing typologies relevant to SRHR discourse in Uganda

o Create and test prompts for commercial LLM APIs (e.g., Claude, GPT-4) to classify content

o Validate outputs against human review on a sample

o Document prompt design, error rates, and limitations

4. Analyze outputs and develop recommendations

o Synthesize findings on sentiment trends and narrative patterns

o Identify shifts associated with policy events (if detectable in sample)

o Translate findings into recommendations for advocacy and communications strategy

Deliverables

1. Data management protocol and analysis framework: Methodology covering: data sources and sampling rationale; collection, cleaning, and preprocessing procedures; classification framework and prompt design for LLM-based analysis; validation approach; ethical and data protection considerations.

2. Metadata summary describing final dataset including sources, inclusion/exclusion criteria, time period, sample size, and any data quality issues.

3. Analytical report summarizing sentiment trends, narrative typologies, and shifts over time, including visualization outputs (e.g., dashboards, charts, or maps).

Experience and skills required

* Must be a Ugandan national, with demonstrated familiarity with the country's media landscape.

* Must have 8-15 years professional experience in media research or data analysis.

* Experience (minimum 3 years) in media research, discourse analysis, or text analytics

* Demonstrated ability to design and implement sentiment analysis, topic modelling, and narrative classification frameworks using AI or machine learning approaches.

* Experience preparing unstructured text data for analysis, including cleaning, deduplication, handling encoding issues, and standardizing formats for LLM-based or computational text analysis

* Working knowledge of LLM-based text analysis (prompt design, output validation).

* Experience integrating quantitative and qualitative analytics - combining computational outputs with discourse or framing analysis - to interpret complex media narratives.

* Strong track record of collaborating with interdisciplinary teams, including data scientists, public health experts, and advocacy organizations.

* Ability to package findings into summary reports and graphics that convey key findings.

* Excellent communication skills (written and oral) in English; proficiency in one or more local languages is a strong asset.

How to Apply

To apply, submit the following:

* CV

* Short summary (maximum one page, bulleted) of your proposed approach to data collection

* Daily rate

* Names and contact information (phone and email) for three references