Glossary of terms
- Binomial model: A statistical model used to analyse data with two possible outcomes, typically denoted as "success" and "failure."
- Covariate: A variable that is not the primary focus of a study but is considered alongside the main variables to understand potential relationships or influences.
- Correlation: A statistical measure that indicates the degree to which two variables change together, often expressed as a coefficient between -1 and 1.
- Data noisiness: The presence of random or irrelevant variations in data, which can make it challenging to extract meaningful information.
- Design-based method: A statistical approach that considers the specific design or structure of a study when making inferences from data.
- DHS (Demographic and Health Survey): A large-scale survey programme that collects data on demographic and health indicators in various countries.
- Elimination: The reduction or eradication of a disease or condition from a specific area or population.
- Friction surface: A spatial model representing the resistance or cost of moving between different locations in a geographic area.
- Geoconnect ID: An identifier used to link or connect geospatial data to specific locations or regions.
- Georeferenced data: Data that are associated with specific geographic coordinates, allowing them to be mapped and analysed in a geographic context.
- Geospatial: Refers to data, information, or activities that have a geographic component and can be represented on a map.
- GLMM (Generalised Linear Mixed Model): A statistical model that extends the generalised linear model by incorporating random effects to account for correlation or hierarchical data structures.
- GTMP (Global Trachoma Mapping Project): A global initiative that ran between 2012-2015 aimed at mapping the prevalence of trachoma, a neglected tropical disease that causes blindness.
- Ground-truthing: The process of verifying or validating data, typically by collecting on-the-ground information to confirm their accuracy.
- ITI (International Trachoma Initiative): An organisation dedicated to the elimination of trachoma as a public health problem.
- Landcover: The physical and biological characteristics of the Earth's surface, including vegetation, water bodies, and built environments.
- Linear relationship: A statistical relationship between two variables that can be represented by a straight line on a scatterplot.
- Logistic regression: A statistical model used to analyse the relationship between a binary dependent variable and one or more independent variables.
- MBG (Model-Based Geostatistics): A statistical approach that combines spatial modelling and geostatistics to make predictions about data in unsampled locations.
- Model calibration: The process of adjusting model parameters to ensure that the model's predictions align with observed data.
- Model distribution: The probability distribution used to describe the variability of data in a statistical model.
- Model validation: The process of assessing how well a statistical or predictive model performs by comparing its output to real-world data.
- Multivariate: Involving multiple variables or factors, often used in statistical analysis to understand complex relationships.
- NTD (Neglected Tropical Disease): A group of infectious diseases that primarily affect people in tropical and subtropical regions, often with limited access to healthcare.
- PBT (Probability of being Below Threshold): A metric used to estimate the probability that the disease elimination target has been achieved.
- Population density: The number of people living per unit area, often measured in individuals per square kilometre or square mile.
- Population weighting: A technique used in statistical analysis to account for variations in population size when making inferences.
- R software: An open-source programming language and software environment commonly used for statistical analysis and data visualisation.
- Random effect: In a statistical model, a variable that represents random or unexplained variation, often accounting for correlation within hierarchical data.
- Raster file: A data format that divides geographic information into a grid of cells, where each cell has a value representing a particular attribute.
- Shapefile: A common file format for storing geographic vector data, such as points, lines, and polygons.
- Variance of the spatial process: A measure of how data points vary in space, indicating the degree of spatial autocorrelation.
- Simulation study: A research method that uses computer-generated data to model and analyse real-world phenomena.
- Spatial confounding: A potential issue in spatial analysis where unaccounted factors lead to misleading results.
- Statistically significant: A result in a statistical analysis that is unlikely to have occurred by random chance, often expressed with a significance level (e.g., p-value).
- Strata: Divisions or subsets used to group data for analysis, often based on specific criteria or characteristics.
- TF (Trachomatous inflammation—follicular): A clinical sign of trachoma, indicating inflammation with follicles on the inside of the upper eyelid.
- Threshold: A specific value or condition used to categorise or make decisions in data analysis.
- TT (Trachomatous Trichiasis): A late stage of trachoma, characterised by the in-turning of the eyelashes that tough the eyeball, which can lead to blindness.
- Univariate: Involving a single variable or factor, often used in statistical analysis to describe or analyse data with only one dimension.