Analysis and Downscaling of Precipitation over East Africa and Djibouti: Observed Data, GCM-CMIP6, and CORDEX

This study provides a multi-scale comparison of simulated and observed precipitation. Global simulations from the CanESM5 model (CMIP6, 282 km) are contrasted with results obtained through stochastic downscaling at 3.5 km using CSTools. Observed rainfall for 1981 is spatially interpolated using Inverse Distance Weighting (IDW).

In addition, climate projections from the EC-Earth3-Veg model (CMIP6) under the SSP585 scenario are analyzed for the period 2021–2040, focusing on both the Republic of Djibouti and the wider East Africa region. Finally, downscaled daily precipitation from CORDEX (1981–1985) is generated using Nearest Neighbor and Bilinear interpolation, allowing an assessment of the sensitivity of results to methodological choices.

Figure 0. Comparison of rainfall variability from satellite products versus observation in situ from 1980 to 2021.

Figure 1. CMIP6 GCM CanESM5 precipitation for 1981 (spatial resolution: 282 km).

Figure 2. CMIP6 GCM CanESM5 precipitation for 1981 downscaled to 3.5 km using stochastic methods with CSTools.

Figure 3. Spatial distribution of observed rainfall in 1981 using Inverse Distance Weighting (IDW) interpolation

Figure 4. Projected total monthly precipitation (mm) from the EC-Earth3-Veg model (GCM-CMIP6), based on the ssp585 scenario Over the Republic of Djibouti during 2021-2040.

Figure 5. Projected total monthly precipitation (mm) from the EC-Earth3-Veg model (GCM-CMIP6), based on the ssp585 scenario Over the Eastern of Africa during 2021-2040.

Figure 6. Downscaled daily precipitation from CORDEX (1981–1985) using (a) Nearest Neighbor interpolation, (b) Bilinear interpolation, and (c) original CORDEX data for comparison.

Figure 7. Performance comparison of the occurrence, duration and intensity of rainfall simulated by Canadian global and regional climate models against the observed rainfall at the airport station.

Abdi-Basid ADAN, 2023

Predictive Analysis of Customer Behavior in E-Commerce: Prediction of Average Order Value and Identification of High-Value Customers

Description:

This data analysis project explores customer behavior on an e-commerce platform using a dataset containing key metrics such as session duration, product detail views, app transactions, add-to-cart rate per session, discount rate per visited products, credit card info saving, average order value (“avg order value”), and a high-value customer indicator (“high_value_customer”). The code is structured in several steps:

Data Preparation: Loading from the clipboard, cleaning (replacing commas with periods for decimals), numeric conversion, and encoding of categorical variables (e.g., yes/no via LabelEncoder). 

Regression Modeling: Use of an XGBoost model to predict average order value, with evaluation via RMSE and R² on a test set (30% of the data). Visualizations include a scatter plot of predictions vs. actual values, a correlation matrix, a boxplot of average basket by card saving, and a histogram of prediction errors. 

Classification Modeling: Logistic regression with L2 regularization to identify high-value customers, based on selected features (session duration, product views, etc.). Evaluation via ROC-AUC score and ROC curve.

Abdi-Basid ADAN, 09–2025

🎯 The detailed methodology and results can be accessed through this link:

👉Click here now! : https://github.com/abdibasidadan-byte

Climatic and Environmental Drivers of NDVI Dynamics in an Arid Region: Predictor Importance, Partial Correlation, and Trend Analysis (1987–2022)

Understanding changes in mangrove ecosystems driven by human activities, climate change, and environmental variations is essential for effective ecological management. This study focuses on the spatiotemporal variability of the Normalized Difference Vegetation Index (NDVI) and examines its responses to parameters such as sea level (SL), Potential Evapotranspiration (PET), rainfall (RF), Standardized Precipitation Index (SPI-1 month), soil moisture (SM), minimum temperature (TN), and maximum temperature (TX) within the study area. Trends, relative influences, spatial autocorrelation, and relationships between NDVI and climatic-environmental variables, as well as partial correlations, were analyzed using the Mann-Kendall monotonic trend test (MKMT), Relative Weight Analysis (RWA), partial correlation coefficients (PCC), and Multiple Linear Regression (MLR) methods.The spatiotemporal patterns of NDVI reveal a reduction in bare soil and an increase in sparse and dense vegetation from 1987 to 2022. Nevertheless, zones of degradation were observed, particularly in southern Godoria in 2022 compared to 1987, as indicated by NDVI. A notable deterioration in NDVI (> 0.2) was recorded from 2000 to 2012, while the overall interannual trend shows a slight decline.Additionally, analyses with Mann-Kendall and Theil-Sen slope reveal that TN, TX, PET, and SPI-1 show increasing trends, though not statistically significant, while SM and LST show decreasing trends. For environmental variables, SL indicates an upward trend. Further, partial correlation analysis identifies SL, TN, SPI-1, TX, and PET as the primary climatic factors controlling vegetation dynamics during the JJAS season, with PCC values of -0.89, 0.87, 0.77, -0.76, -0.75, and 0.86 with NDVI, respectively.These findings highlight the significant influence of select environmental factors on the spatiotemporal dynamics of mangrove vegetation, providing insights critical for conservation and management efforts.

 

2.1.1.  Partial correlation

The bivariate correlation coefficients may not effectively represent the complex relationships among variables in multivariate correlation analysis, given that multiple factors can influence these relationships. Therefore, partial correlation coefficients were computed to assess the spatiotemporal strength and direction of the linear relationship between NDVI and each climate variable, while controlling for the effects of the other climate variables (i.e, sea level, PET, SM, SPI, LST, TN and TX). The strongest correlation is close to 1, while the weakest is below 0.5. Thus, the partial correlation can be calculated as follow (Cheng et al., 2017):

                                                                                                 

(1)

Where x, y, and z represent three distinct variables. R_xy,z signifies the partial correlation between variables x and y while accounting for the influence of variable z. Similarly, R_xy denotes the linear correlation coefficient between x and y, with R_xz​ and R_yz​ conveying analogous interpretations.

To explore the spatial autocorrelation of NDVI data, we used the "global and local autocorrelation analysis based on Moran’s I statistics." This method enables the evaluation of the average spatial differences between individual cells and their adjacent neighbors, thereby characterizing the spatial attributes of a specific property across the entire study area through global spatial autocorrelation analysis. In Moran’s statistics, the normalized z-score can range from -1 to +1. A Moran's I value exceeding 0 indicates a positive correlation, which suggests a clustering pattern, while a value below 0 points to a negative correlation, reflecting a dispersed arrangement. The calculation of Moran’s I statistics for examining spatial autocorrelation is provided by Xu et al. (2015):

 (2)

Where N represents the number of observations, x_i​ denotes the observed value for cell i, x_j ​ indicates the observed value for cell j,  is the average of cell i or cell j, and w_ij​ is the weight assigned to the relationship between cells i and j.

While the global spatial autocorrelation through Moran's I statistics reveals the overall clustering pattern, it does not allow for the assessment of spatial association patterns across multiple locations. In contrast, Local Spatial Autocorrelation focuses on the significance of local statistics at each individual location and identifies the presence of spatial clusters, a capability that global spatial autocorrelation lacks. The mathematical equation of local spatial autocorrelation using Moran's I is described by (Anselin (2010)).

 (3)

(a)   (b)

Fig. 1. (a)Estimation of relative importance of climate variables as predictors of NDVI. (b) Pearson correlation coefficient between NDVI and climate variables.

 

 

Table 1. Mann Kendall Trend and Theil-sen slope statistics of NDVI and climate variables from 1987 to 2022.

	Theil-sen slope	p.value	Mann kendall tau (τ)	p.value
NDVI	-0.001	0.858	-0.023	0.859
Sea level	0.001	0.646	0.057	0.646
Rainfall	0.000	NA	0.000	NA
SPI	0.000	0.917	0.000	0.917
Temp Min	0.057	0.109	0.194	0.109
Temp Max	0.072	0.173	0.166	0.172
Soil Moisture	-0.030	0.739	-0.057	0.739
PET	0.046	0.649	0.085	0.649
LST	-0.222	0.720	-0.111	0.720

 

Abdi-Basid ADAN, 2024

The Abdi-Basid Courses Institute (TABCI)