As we need to analyze the suitability of an area based on some preferences, we are basically doing an MCDA (Multi-criteria decision analysis). MCDA, in GIS, is generally done with raster data, and the final map shows the suitability of every cell in the study area. We can use MCDA for different purposes, like analyzing the suitability of the land for a specific species, or choosing the right site for a building with quantitative needs. During the process, we have to create raster maps for every criteria, then calculate the final suitability based on them. For this task, we differentiate between these two kinds of data:

• Constraint: Binary raster maps having cells with the value of zero (not suitable for the task), and having cells with the value of one (suitable for the task). These binary raster layers can be considered as masks, and define the areas we can classify in our final assessment.
• Factor: Raster maps showing the possibility that a cell will be suitable for a given criteria, also called fuzzy maps. Their values are floating point numbers between 0 and 1 (0 represents 0%--absolutely sure it is not suitable, while 1 represents 100% --absolutely sure the cell is suitable).

In the end, we will have to create a single map by combining the different constraints and factors, showing the overall suitability of the cells calculated from the different factors, and masked by the union of the different constraints. There are several approaches and steps to execute an MCDA analysis, although in GIS, the most popular approach is to use the multi-criteria evaluation (MCE) method. By using this method alone, the result will have some uncertainty due to the involved subjectivity, although it will suit us in our task. First, let's break down our criteria to constraints and factors as follows:

• Constraints: Study area, maximum 5 kilometers from main roads, specific land use types, slope less than 10 degrees, minimum 200 meters away from waterways and water bodies
• Factors: Close to main roads, close to the mean point of the appropriate settlements, far from waterways and water bodies

Using this naive grouping, we have to process some of our data twice, as we have some overlaps between our constraints and our factors. However, we do not need to use those data as both constraints and factors. We can normalize our factors in a way that the constrained areas automatically get excluded from the result. Furthermore, as our DEM is already clipped to the borders of our study area, we do not have to create a raster layer from our study area. That is, we can regroup our tasks in the following way:

• Constraints: Specific land use types, slope less than 10 degrees
• Factors: Close to main roads (maximum 5 kilometers), close to the mean point of the appropriate settlements, far from waterways and water bodies (minimum 200 meters)