Water is an important basic need for human survival. Many Americans obtain water from public water systems, however, about 45 million Americans use private wells for drinking water. When water is contaminated, it becomes unsafe for consumption and can cause many poor health outcomes. As a result, many environmental hazards present in water are regulated in public water systems, however, private well owners are not required to test, disinfect or treat their water. In Gaston County, North Carolina, private wells provide the primary water supply for approximately 42% of the residents. Since 1989, well permits issued for new wells have been stored primarily on paper. Lack of digitization has hindered the ability of researchers and public health officials to access private well information. Further, lack of well testing data including arsenic and coliform has also made it difficult to determine groundwater quality in wells. No studies exist that describe spatial variation of arsenic and coliform bacteria presence in wells in Gaston County, that is within the NC Piedmont geologic belt.The main objective of the dissertation is to incorporate multilevel geocoding and spatial modeling techniques to predict the risk of arsenic and coliform bacteria in private wells. To achieve this goal, first a GIS database of private wells is created using geocoding. Because the positional accuracy of private wells in GIS can affect the spatial analysis results, global positioning system (GPS) coordinates were obtained at 1035 wells to compare differences in the results of rooftop, parcel and street geocoding techniques. Second, the multilevel geocoding approach was used to determine the geographic coordinates of arsenic samples from 2011 to 2017. The sampled arsenic information, data on geology, pH, and well depth were used to estimate the probability of having arsenic at or above detectable levels (≥ 5 µg/L) in wells across the county. This threshold was used because low levels of arsenic, even below the drinking water standard of 10 µg/L set by the United States Environmental Protection Agency, are still detrimental to human health and most of the arsenic detections in the study are between 5-10 µg/L. Third, coliform samples from private wells, well characteristics, parcel size, and soil ratings for the leachfield are examined to estimate the probability of having coliform bacteria in a well. A multilevel geocoding approach improved match rate of permit addresses from 38.0% (n = 3,318) to 98.9% (n = 8,616). Addresses that were re-engineered during geocoding accounted for 50.9% (n = 4,439) of the matched records in the GIS database. There were significant differences (p < 0.05) in positional accuracy for rooftop, parcel, and street geocodes of private wells in the GIS database; positional accuracy was highest for rooftop geocodes. Private wells set in mica schist (ꞒZms) were associated with arsenic at detectable levels, suggesting a local-scale geologic source influence of arsenic in the county. In addition, pH (median = 7.1) was positively associated with the presence of arsenic in well water, indicating arsenic ≥ 5 µg/L was predominantly associated with pH > 7.3. An area in the northwestern section (8.4 km2) of the county was identified as having more than 50 percent likelihood of arsenic concentrations ≥ 5 µg/L. This area was found in the inner Piedmont of North Carolina belt and coincide with the mica schist geologic rock. The multivariate logistic regression model results indicate that bored and older wells are more likely to have a high probability of coliform bacteria. The lack of significant association between poorly rated soils for a leachfield and probability of having coliform bacteria suggests that contamination is not a result of pathogens in household wastewater. There was no association between well depth and probability of having coliform in a well suggesting that contamination may come from runoff water.Overall, the advanced geocoding approach can be used to improve geocoding match rate of input addresses for analytical purposes and develop a GIS database of private wells. The analysis of arsenic data in combination with geology, well depth and pH can provide preliminary insight into causes of long-term exposure to arsenic in groundwater. There was a higher chance of finding coliform bacteria in bored older wells. Because older wells (average well age = 19 years) were significantly likely to contain coliform bacteria suggest that those constructed before well standards was enforced may have a higher issue with coliform bacteria. GIS maps can now be leveraged for targeted interventions to affected private wells in Gaston County. The present study is applicable to other regions interested in developing a GIS database of private wells, and towards advance understanding of spatial analysis of water hazards when few samples are taken in the field. This study provides a holistic approach that can be adopted for other regions facing similar groundwater exposure to environmental hazards in private wells.