Research Projects
Project Guidelines
These guidelines are designed to help you conduct your research
projects using the spatial analysis techniques learned in this course.
Each project will follow a similar structure, allowing you to apply the
concepts and methods consistently.
General Steps for Each Project:
- Operationalize the Outcome Variable:
- Clearly define your outcome variable (the problem you are
investigating).
- Explain how you will measure this variable using the available
data.
- Provide the specific variable name(s) from the dataset that you will
use.
- Conceptualize Covariates:
- Identify the explanatory variables (covariates, factors, or
determinants) that you will use in your analysis.
- Provide a theoretical or literature-based justification for why
these variables are relevant.
- Cite leading paper(s) that support your choice of covariates.
- Specify the variable names from the dataset that you will use.
- Data Selection and Cleaning:
- Data Source: Specify the dataset you will use
(e.g., DHS, Integrated Household Budget Survey, Labour Force
Survey).
- Spatial Data: Identify and load the appropriate
shapefile for your analysis.
- Multilevel Data: Load the survey data and prepare
it for analysis.
- Data Cleaning:
- Handle missing values appropriately.
- Recode variables as needed.
- Ensure all necessary variables are present and correctly
formatted.
- Create new variables as needed.
- Spatial Analysis:
- Confirmation (Statistical Tests):
- Calculate Global Moran’s I and its p-value to determine if spatial
analysis is appropriate for your outcome variable.
- Interpret the results.
- Spatial Autocorrelation:
- Global Moran’s I: Assess the overall spatial
clustering of your outcome variable.
- Local Moran’s I: Identify regions/districts that
are correlated in terms of your outcome variable.
- Spatial Distribution:
- If the p-value of Global Moran’s I is less than 0.05, visualize the
spatial distribution of your outcome variable (proportion).
- Hot Spot and Cold Spot Analysis:
- Calculate and interpret the Getis-Ord Gi* statistic to identify hot
spots and cold spots.
- Univariate Analysis:
- Categorical Variables: Calculate frequencies and
percentages.
- Continuous Variables: Calculate means and standard
deviations.
- Bivariate Analysis:
- Association: Use chi-square tests for categorical
independent and dependent variables.
- Effects: Use ANOVA for categorical independent
variables and continuous dependent variables.
- Dependency: Use covariance to assess the
relationship between continuous variables.
- Multivariable Analysis:
- Regression Models: Use only variables that were
significant in the bivariate analysis.
- Specify the type of regression model used based on the outcome
variable (e.g., linear, logistic, Poisson).
- Survey: Hierarchical Data Sets: Multilevel
Analysis:
- Model 0: Include only the dependent variable.
- Model 1: Include the dependent variable and
individual-level factors.
- Model 2: Include the dependent variable and
household-level factors.
- Model 3: Include the dependent variable and
community-level factors.
- Model 4: Include the dependent variable and all
levels.
- Model Comparison:
- Model Adequacy:
- Calculate the Intra-class Correlation (ICC). If ICC > 0%,
continue to multilevel analysis.
- Model Evaluation:
- Use AIC, BIC, LRT, or other appropriate model comparison
metrics.
Project Assignments:
- Project 1 (Online Campus Students):
- Topic: Spatial distribution and Determinants of
Unprotected Water Source among Households in Somalia: Spatial and
Multilevel Analysis using Nationwide Survey Data.
- Data: Use the DHS data for Somalia.
- Project 2 (Hargeisa Campus Students):
- Topic: Spatial Patterns and Factors Associated of
Unskilled Birth Attendance among Reproductive Age Women in Somaliland:
Spatial and Multilevel Analysis using Nationwide Survey Data.
- Data: Use the DHS data for Somaliland.
- Project 3 (Borama Campus Students):
- Topic: Spatial Variation and Determinants of School
Attendance among Adults in Somalia: Spatial and Multilevel Analysis
using Somalia Integrated Household Budget Survey.
- Data: Use the Somalia Integrated Household Budget
Survey.
R Markdown Template:
Sample of Your Document
Introduction
[Provide a brief introduction to your research topic, its importance,
and the research questions you will address. Briefly mention the data
source and methods.]
Data and Methods
Libraries
# Load necessary R libraries
library(sf)
library(ggplot2)
library(spdep)
library(tmap)
library(haven)
library(survey)
library(RColorBrewer)
Data Loading and Cleaning
# Load shapefile
# Specify the file path to your shapefile
#filepath <- "path/to/your/shapefile.shp"
#shapefile <- st_read(filepath)
# Load survey data
# Specify the file path to your survey data (.dta file)
#surveydata <- read_dta("path/to/your/surveydata.dta")
# Data cleaning and pre-processing steps
# ... (add your data cleaning code here)
Operationalizing Outcome Variable
[Clearly define your outcome variable and how you will measure it.
Provide the variable name(s) from the dataset.]
Conceptualizing Covariates
[Identify your explanatory variables, provide a theoretical
justification, cite leading papers, and specify variable names.]
Spatial Analysis
Spatial Weights Matrix
# Create spatial weights matrix
# ... (add your code to create spatial weights matrix)
Global Moran’s I
# Calculate Global Moran's I
# ... (add your code to calculate Global Moran's I)
Local Moran’s I
# Calculate Local Moran's I
# ... (add your code to calculate Local Moran's I)
Getis-Ord Gi*
# Calculate Getis-Ord Gi*
# ... (add your code to calculate Getis-Ord Gi*)
Spatial Visualization
# Create maps using tmap or ggplot2
# ... (add your code to create spatial maps)
Statistical Analysis
Univariate Analysis
# Perform univariate analysis
# ... (add your code for univariate analysis)
Bivariate Analysis
# Perform bivariate analysis
# ... (add your code for bivariate analysis)
Multivariable Analysis
# Perform multivariable analysis
# ... (add your code for multivariable analysis)
Multilevel Analysis
# Perform multilevel analysis
# ... (add your code for multilevel analysis)
Model Comparison
# Perform model comparison
# ... (add your code for model comparison)
Conclusion
[Summarize your findings, discuss their implications, and suggest
future research directions.]
References
[Include a list of references cited in your report.] ```
How to Use These Guidelines:
- Choose Your Project: Select the project assigned to
your campus.
- Review the Guidelines: Carefully read the general
steps and the specific requirements for your project.
- Use the R Markdown Template: Copy the R Markdown
template and start filling in the sections with your analysis, code, and
interpretations.
- Follow the Steps: Work through each step of the
guidelines, ensuring that you address all the required components.
- Write Up Your Results: Write a clear and concise
report that explains your methods, results, and conclusions.
- Prepare for Presentation: Be ready to present your
findings to the class.
These guidelines and the template should provide a solid foundation
for your research projects. If you have any questions, please ask!