Monitoring School Meal Quality - 2

Describe the structure, scale, and nutritional goals of India's Mid-Day Meal Scheme (PM POSHAN).

Identify common categories of meal quality violations (e.g., missing food groups, contamination) from published reports or news sources.

Describe how meal quality monitoring currently works in practice and identify its key limitations.

Map the project's available data sources (images, text, metadata) to the specific types of information each can provide.

Outline a data collection plan, specifying target sources, acquisition methods, and ethical considerations.

Formulate a precise, measurable definition of a 'violation' that can be operationalized using the available data.

Classify each violation type by the data modality (image, text, metadata) most likely to capture evidence for it.

Decompose the overall project goal into concrete ML sub-tasks (e.g., image classification, text classification) with specified inputs and outputs.

Identify gaps where no available modality provides a strong signal for a given violation type, and propose strategies to handle them.

State the key assumptions and limitations of framing meal quality monitoring as a set of ML tasks.

Design a structured, multi-label annotation schema for meal images that captures quality, quantity, and variety.

Design an annotation schema for complaint text that captures violation type and severity level.

Write clear, unambiguous labeling guidelines for each schema that a non-expert annotator could follow.

Explain the importance of inter-annotator reliability and define a strategy for resolving disagreements during labeling.

Plan a small pilot annotation exercise to test and refine the labeling guidelines on sample data.

Identify common types of noise and irrelevance in scraped social media data for this project.

Define and apply filtering criteria to remove low-quality or off-topic samples from both text and image data.