Data: sources, preparation and manipulation

My main source of data was found on the S drive. After continually searching the web for data on Kenya I found nothing comparable to what was on the S drive. I derived data from both the Kenya folder and the Africa folder. Almost all data sets needed to be prepared in order to use and manipulate.

Preparation of the data consisted of digitizing rubber sheeting and some calculations. I digitized the parks and reserves adding missing parks that I felt were important to the study based on area size. Small parks under 100 square km where not included. I found the park shape and location on a number of different web sites. I use more than one site because there wasn't one single comprehensive site.

Rubber sheeting was another technique I used to prepare my data. The data found in the Africa folder had a different projection than that used with the Kenya data. Also the data was for all of Africa rather than just Kenya. The Africa folder data had a projection of lon/lat and the Kenya data was plane. Other methods could have been used but for reasons that will be explained in the problems section of the report. In order to rubber sheet the Africa layers into the same format as the Kenya layers I had to collect a number of points coordinates that were of the same location on both the Africa layers and Kenya layers. I selected all of my points along political boundaries because they provided sharp corner, which were easily matched from one layer to the next. Also my data provided me with no other reference points such as cities or water bodies that were common to both layer types. Once the coordinates were found a correspondence file was made in edit. The coordinates were matched, the old with the new. In the end this operation provided a new layer for each of the Africa layers being used in the analysis. These layers could now be overlaid with the Kenya layer.

The calculations that were included in my data preparation were performed on the Kenya rainfall layer. This layer originally provided 1-3 different values for the type of rainfall within a given year for each zone. The values were given on a scale of Humid to Dry. Humid being 0 and Dry being 5. Then there were percentages given to the 1-3 different values for the amount of time each corresponding rainfall type was found throughout the year. These values would make it complicated to make any analysis so a new value had to be calculated. Setting the scale from 1-6 did this, Humid being 1 and 6 being Dry. The values for each zone were multiplied by the corresponding percentages. These were then adding to give a final single value. The smaller the value the more humid or the greater the amount of rainfall. After assigning the new values to the layer I checked the new display with the old to confirm that my calculations provide the same distribution of rainfall.

One of the main components in the final steps of my analysis is concavity index. This was used to analyze the perimeter to area ratio. The area in cells and perimeter in km was found by performing the area and per under the analysis heading. Then by using the formula (perimeter)*(number of cells^ -1/2) a value was given. This value ranged from 23.14 to 53.75. The lowest number being most like a circle or the smallest perimeter to area ratio which is most desirable.