Table 2.
Results of Quasi Poisson Generalized Linear Regression Models with and without Spatial Eigenvector Mapping for cumulative number of buildings or highways added to OSM in regard to to overall and humanitarian mapping activities.
| All | Humanitarian | |||||
|---|---|---|---|---|---|---|
| Estimate | Std. err. | z value | Estimate | Std. err. | z value | |
| Buildings, SEVM | ||||||
| (Intercept) | − 12.33*** | 0.62 | − 19.79 | − 15.62*** | 1.22 | − 12.77 |
| log(population) | 13.24*** | 0.33 | 40.37 | 13.75*** | 0.64 | 21.56 |
| shdi | 0.41*** | 0.04 | 10.83 | − 1.13*** | 0.08 | − 14.29 |
| cellphone users | 0.11** | 0.03 | 3.26 | 0.26*** | 0.06 | 4.06 |
| gender equality | 0.21*** | 0.03 | 7.30 | 0.15** | 0.06 | 2.62 |
| disaster deaths | 0.39*** | 0.04 | 9.01 | 0.28*** | 0.05 | 5.51 |
| youthmappers | 0.03** | 0.01 | 2.77 | 0.04* | 0.02 | 2.16 |
| explained deviance: 0.79 | explained deviance: 0.68 | |||||
| Eigenvectors: 12 | Eigenvectors: 8 | |||||
| Buildings, without SEVM | ||||||
| (Intercept) | − 10.93*** | 0.77 | − 14.12 | − 17.23*** | 1.20 | − 14.40 |
| log(population) | 12.56*** | 0.40 | 31.02 | 14.79*** | 0.62 | 23.88 |
| SHDI | 0.62*** | 0.05 | 13.72 | − 1.27*** | 0.07 | − 17.56 |
| cellphone users | − 0.003 | 0.42 | − 0.09 | 0.27*** | 0.05 | 5.14 |
| gender equality | 0.38*** | 0.03 | 11.29 | 0.34*** | 0.05 | 7.19 |
| disaster deaths | 0.58*** | 0.05 | 12.17 | 0.19*** | 0.05 | 3.57 |
| youthmappers | − 0.007 | 0.02 | − 0.43 | 0.03 | 0.02 | 1.62 |
| explained deviance: 0.60 | explained deviance: 0.54 | |||||
| Highways, SEVM | ||||||
| (Intercept) | − 11.78*** | 0.38 | − 31.11 | − 20.06*** | 1.43 | − 13.99 |
| log(population) | 12.32*** | 0.20 | 61.44 | 14.54*** | 0.74 | 19.65 |
| SHDI | 1.03*** | 0.03 | 35.46 | − 1.29*** | 0.09 | − 14.60 |
| cellphone users | 0.04* | 0.02 | 2.04 | 0.03 | 0.07 | 0.056 |
| gender equality | 0.12*** | 0.02 | 7.32 | 0.15* | 0.06 | 2.52 |
| disaster deaths | 0.21*** | 0.05 | 4.32 | 0.37*** | 0.05 | 7.17 |
| youthmappers | − 0.01 | 0.001 | − 1.05 | 0.04 | 0.02 | 1.86 |
| explained deviance: 0.91 | explained deviance: 0.72 | |||||
| Eigenvectors: 16 | Eigenvectors: 7 | |||||
| Highways, without SEVM | ||||||
| (Intercept) | − 9.56*** | 0.43 | − 22.35 | − 23.11*** | 1.54 | − 15.00 |
| log(population) | 11.10*** | 0.22 | 49.49 | 16.33*** | 0.79 | 20.65 |
| SHDI | 1.10*** | 0.03 | 36.21 | − 1.56*** | 0.10 | − 16.20 |
| cellphone users | 0.07** | 0.03 | 2.74 | 0.11 | 0.07 | 1.64 |
| gender equality | 0.14*** | 0.02 | 7.21 | 0.46*** | 0.06 | 8.18 |
| disaster deaths | 0.32*** | 0.05 | 6.34 | 0.28*** | 0.06 | 4.76 |
| youthmappers | − 0.02* | 0.01 | − 2.07 | 0.003 | 0.03 | 0.14 |
| explained deviance: 0.83 | explained deviance: 0.60 | |||||
The spatial eigenvectors contribute to the explained deviance. For models without incorporation of spatial autocorrelation standard errors are too small, coefficients might be biased - these models are only reported for comparison. Regression coefficients are reported for standardized predictors to make the effect of the different predictors comparable. The only exception is population which has not been standardized due to the use of the log-transformation and since population is included as a covariate that needs to be controlled for but not of primary interest. Significance levels: .