Upon completion of this session and engagement with this resource, you will be able to:

• Define fundamental data concepts (e.g., dataset, variable, data types, big data) in the context of public health.
• Identify key sources of public health data and recognize indicators of data quality.
• Explain the basic principles of data visualization and its importance in communicating public health information.
• Describe the core concepts of Artificial Intelligence (AI), Machine Learning (ML), and Large Language Models (LLMs) without technical jargon.
• Utilize prompt engineering techniques to effectively interact with LLMs (like ChatGPT/Gemini) for tasks relevant to public health.
• Recognize common open-source software and AI tools accessible to non-coders for basic data exploration and AI interaction.
• Discuss ethical considerations and potential biases associated with using data and AI in public health.
• Apply these concepts through practical, non-coding exercises simulating real-world public health scenarios.

• The "Why": Why data and AI/ML are no longer just for tech experts but essential for all healthcare professionals.
  • Example: How rapid data analysis helped track COVID-19 spread and inform public response.
• What We'll Cover: A roadmap of the session, emphasizing the non-coding, practical approach.
  • Understanding the Language of Data
  • Finding and Trusting Your Data
  • Making Data Speak: The Power of Visuals
  • Demystifying AI and Your New AI Assistant (LLMs)
  • Using Data and AI Responsibly
• Focus on Empowerment: This session is designed to give you the confidence to engage with data and AI tools, improving your daily work and contributions to public health.

• Tools: ChatGPT (OpenAI) / Gemini (Google) / Claude (Anthropic)
  • Main Features: Natural language understanding and generation, summarization, brainstorming, drafting text, answering questions.
  • Practical Healthcare Applications:
    • Drafting patient education materials (e.g., "Explain type 2 diabetes in simple terms for a brochure").
    • Summarizing research papers or public health reports for quick insights.
    • Brainstorming public health campaign slogans or outreach strategies.
    • Generating FAQs for common health concerns.
  • Prompt Engineering Focus:
    • Concept: How to ask the right questions to get the best results. Explain "Role, Task, Format, Constraints."
    • Example:
      • Weak Prompt: "Tell me about vaccination."
      • Strong Prompt: "Act as a public health advisor. Create a list of 5 key benefits of childhood vaccination for parents of newborns, written in clear, empathetic language, suitable for a pamphlet. Each benefit should be one sentence."
• Tools: PubMed Central / Perplexity AI / Semantic Scholar
  • Main Features: Conversational search engine that provides answers with cited sources. Excellent for research and fact-finding.
  • Practical Healthcare Applications:
    • Quickly finding evidence-based information on specific health conditions or interventions.
    • Identifying recent research papers on a public health topic.
    • Checking the source of health claims.

• Tools: Google Forms / Microsoft Forms
  • Main Features: Easy-to-create surveys, quizzes, and feedback forms. Collects responses in a spreadsheet.
  • Practical Healthcare Applications: Community health needs assessments, patient satisfaction surveys, collecting sign-ups for health workshops, post-event feedback.
• Tool: KoboToolbox
  • Main Features: Free, open-source suite for field data collection, often used in humanitarian and development contexts. Works offline.
  • Practical Healthcare Applications: Epidemiological surveys in remote areas, health facility assessments, monitoring public health interventions.

• Tools: OpenRefine / Google Sheets / Microsoft Excel (Online/Free versions)
  • Main Features: Organizing data in tables, basic calculations (sum, average), creating simple charts (bar, pie, line).
  • Practical Healthcare Applications: Tracking patient appointments, managing small project budgets, creating simple dashboards for clinic metrics, and visualizing immunization rates over time.
• Tools: DataVoyager / Flourish / Datawrapper
  • Main Features: Web-based tools for creating interactive and embeddable charts, maps, and tables with no coding. Generous free tiers.
  • Practical Healthcare Applications: Creating compelling visuals for public health reports, presentations, or websites (e.g., mapping disease prevalence, showing trends in health behaviors).

• Content: Defining data, datasets, variables, data types (quantitative vs. qualitative). Importance of context. What is "Big Data" in simple terms?
• Experiential Learning Use Cases (Non-Coding):
  1. Scenario Analysis: Given a brief public health scenario (e.g., a local flu outbreak), identify 5 types of data that would be useful to collect (e.g., number of cases, age of patients, vaccination status, onset date, symptoms). Classify each as quantitative or qualitative.
  2. "Spot the Data": Look at a simplified public health infographic (provided). Identify 3 key data points and what they represent.
  3. Data Scavenger Hunt: Find a public health statistic from a reputable source (e.g., WHO, CDC website). Describe what it measures and its unit of measurement.
  4. Dataset Exploration (Conceptual): Review a very small, clean sample dataset (e.g., 10 rows, 5 columns in a table about patient demographics and a health outcome). Identify the variables and their likely data types.

• Content: Common public health data sources (surveys, surveillance systems, electronic health records, census data). Characteristics of good quality data (accuracy, completeness, timeliness, reliability, relevance).
• Experiential Learning Use Cases (Non-Coding):
  1. Source Evaluation: Given two fictional data sources for child malnutrition rates (one from a well-known NGO, one from an anonymous blog), discuss which is likely more reliable and why.
  2. "Is This Data Healthy?": Review a small, sample dataset with obvious errors or missing values (e.g., age = 150, city missing for half the entries). Identify 3 quality issues.
  3. Survey Critique: Review a short sample survey (provided) with leading questions or biased options. Identify 2-3 problematic questions and suggest improvements.
  4. Brainstorming Data Gaps: For a specific public health goal (e.g., reducing smoking in teens), brainstorm what data is needed and where it might be found or how it could be collected ethically.

• Content: Why visualize data? Common chart types (bar, line, pie, map) and when to use them. Principles of clear and honest visualization (avoiding misleading charts).
• Experiential Learning Use Cases (Non-Coding):
  1. Chart Match-Up: Given 3 simple datasets (e.g., disease cases over time, comparison of risk factors by percentage, geographical distribution of clinics) and 3 charts types (line, pie, map), match the best chart to each dataset.
  2. "Bad Chart" Detective: Analyze a misleading chart (provided) and explain why it's problematic (e.g., truncated y-axis, confusing colors).
  3. Sketch-a-Visual: Given a public health message (e.g., "Cases of X disease have increased by 50% in the last year among young adults"), sketch a simple chart idea to communicate this effectively.
  4. Interactive Exploration: Use a link to a pre-made interactive chart on Flourish or Datawrapper (e.g., showing global health indicators). Explore the chart and write down two insights you gained.

• Content: Simple definitions of AI, Machine Learning (learning from data), and Large Language Models (understanding and generating text). Focus on what LLMs can do for them.
• Experiential Learning Use Cases (Non-Coding):
  1. LLM Task Brainstorm: List 3 routine tasks in your public health role (e.g., drafting emails, summarizing meeting notes, finding information) where an LLM like ChatGPT could assist.
  2. Prompt Practice - Summarization: Take a short paragraph from a public health news article (provided). Use ChatGPT/Gemini with the prompt: "Summarize this text in one sentence for a busy public health official." Compare the output to the original.
  3. Prompt Practice - Idea Generation: Use ChatGPT/Gemini with the prompt: "Act as a health communication specialist. Brainstorm 5 catchy slogans for a campaign encouraging handwashing in primary schools."
  4. AI Output Critique: Given a short AI-generated text about a health topic, identify one strength and one potential area for improvement or fact-checking. (e.g., Is it too generic? Does it cite sources? Is the tone appropriate?).

• Content: Importance of data privacy (anonymization, confidentiality). Potential for bias in data and AI algorithms. Fairness, accountability, and transparency in AI.
• Experiential Learning Use Cases (Non-Coding):
  1. Case Study Discussion: Read a short, fictional case study about an AI tool used for disease prediction that shows biased results against a certain demographic. Discuss 2-3 ethical concerns.
  2. Privacy Brainstorm: A local clinic wants to share patient data for a research study on diabetes trends. What are 3 key privacy considerations or steps they must take before sharing? (e.g., de-identification, informed consent).
  3. Bias Identification: "An AI model is trained primarily on health data from one ethnic group to predict heart disease risk." Discuss potential bias if this model is then applied to a diverse population.
  4. "Ethical AI" Checklist Creation: Brainstorm a 3-5 point checklist of questions a public health professional should ask before adopting a new AI tool in their work (e.g., Where did the data come from? How is privacy protected? Has it been tested for fairness?).