ACME International Industries - Site location in Anytown region

ACME International Industries – Site location in Anytown region Jan Stafleu ACME International Industries – Site location in Anytown region Author: Jan Stafleu. Date: January 26, 2003. Word count: 1988. 1 Introduction ACME International Industries builds and delivers technical components to the computer industry. It is a highly successful transnational organization and due to the recent expansion of its business in Europe, ACME has decided to locate its new European Headquarters in the United Kingdom. The headquarters will provide much needed employment for 5000 people in the region. After a review of the regional inward investment awards available it has decided to refine its search for new facilities in the region of Anytown, in the Northwest of England. This report describes how GIS was employed in the search for the ideal location for ACME’s new headquarters. Chapter 2, Data Analysis, describes which data was used in the analysis and why. Chapter 3, Methods, describes how the criteria set by the management of ACME were translated in operational criteria for the analysis. It also provides a cartographic model describing the methodology used. The cartographic model was implemented in Idrisi32, release 2.0. The results of this implementation are presented in Chapter 4. Finally, a discussion on the analysis is given in Chapter 5 along with recommendations for future research. 2 Data analysis Anytown This dataset provides a good base map. Attribute data includes nominal data (index number and name) and ratio data (area and perimeter). We have used this dataset for presentation purposes only (ward boundaries and ward names). Unemployment Unemployment is a relatively recent dataset (2000). It is important, since ACME has 5000 new jobs to fill. Unfortunately, the data are of the ordinal type. This means that we are only able to establish that unemployment in one ward is higher than in another. For example, we know that Brinnley has the highest level of unemployment in Anytown; however, we cannot say that in Romsey, unemployment is twice that of North Marpuley. Jobseekers This is an interesting dataset that provides us with ratio data on a potential workforce. The limitations of the unemployment data do not apply to this dataset: we may now state that Romsey has about twice as much jobseekers as North Marpuley. However, the dataset dates from 1998 and we consider that as too old. The number of Jobseekers depends on the economy, which rapidly changes. For example, if another Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 1 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu company opened a factory in Anytown in 1999, providing 2000 jobs or more, this would not be visible in the data. University This is an interesting dataset that provides us with ratio data giving an indication of education levels in each ward. However, data are from 1998 and considered too old for our analysis. Deprivation and Income These datasets are irrelevant; we are not looking for poor or wealthy wards. Health As ACME is a responsible company, management qualifies the use of health data as unethical. Access Access to services is not a criterion for the analysis. Furthermore, it is not clear what is meant by “services”: are these consumer services or business services? Employees The date of this dataset is unknown (1998?). Quick analysis shows the number of employees in Brinnley exceeds the number of residents in that area. This implies that the data includes employees from outside the ward, or maybe even from outside Anytown. It is therefore of limited use in our analysis. Education This is a relatively recent dataset (2000). It is important, since we need to establish the level of education. However, the dataset has the same drawbacks as Unemployment. Population Population is important for establishing the potential workforce in a ward. The dataset is from 1998, but unlike the economy, demography develops slowly, so this dataset can be used. Transportation Railways are necessary for the transfer of goods and public transport for the workforce. However, there is no point in having a site near a railway if there is no rail station. We have therefore used Rail stations rather than Railways. In addition, the presence of a rail station is an indication of the quality of public transportation. The same applies to Bus stations. Major roads The new premises must be located near major roads. However, since all railway stations are accessible via major roads, this criterion is automatically satisfied if we select a site near a railway station. Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 2 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu 3 Methods 3.1 Operational criteria The criteria set by the management of ACME were translated in the following operational criteria that can be used in our analysis: 1. The site must be in or near wards with a high potential workforce; 2. To provide good public transport, the site must be located within 1000 meters of rail station(s) or major bus station(s); 3. To allow the transfer of goods, the site must be located at or in very close proximity of a rail station and on one or more major roads. Since all rail stations are accessible by major roads, we can limit this criterion to rail stations only; 3.2 Potential Workforce To establish the potential workforce in a ward we have used the datasets Unemployment, Education and Population. Since the first two datasets are ordinal, we have not calculated the actual number of potential workers, but rather established a “potential workforce factor”. This potential workforce factor was calculated as follows: Potential workforce factor = unemployment factor * education factor * population factor. The higher the potential workforce factor, the better the ward is suited for ACME. 3.2.1 Unemployment factor We have divided the unemployment data of Anytown into five classes of equal class width. Each class was assigned an unemployment factor ranging from 0.2 (low unemployment) to 1.0 (high unemployment): Class Lower boundary Upper boundary Factor 1 224 1,635 1.0 2 1,636 3,046 0.8 3 3,047 4,457 0.6 4 4,458 5,868 0.4 5 5,869 7,279 0.2 Thus, the higher the unemployment, the greater the contribution to the potential workforce factor. 3.2.2 Education factor The same procedure was applied to the Education levels in Anytown. The factors, however, are different. Since ACME wants to provide training, we have assigned a low factor to highly educated wards. However, education levels should not be too low either, so we have assigned a low factor to the lowest educated wards as well: Class Lower boundary Upper boundary Factor 1 1,644 1,749 0.2 2 1,750 3,393 1.0 3 3,394 5,037 1.0 4 5,038 6,681 1.0 5 6,682 8,325 0.2 Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 3 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu 3.2.3 Population factor The higher the population, the greater the potential workforce: Class Lower boundary Upper boundary Factor 1 10,200 11,461 0.2 2 11,462 12,722 0.4 3 12,723 13,983 0.6 4 13,984 15,244 0.8 5 15,245 16,505 1.0 3.3 Cartographic model Figure 1 shows the cartographic model used in our analysis. unemployment education population railway_station bus_station RECLASS RECLASS RECLASS POINTRAS POINTRAS unemployment education population railway station bus station reclass reclass reclass raster raster OVERLAY OVERLAY pwf = u * e * p pt = r + b potential public workforce transport BUFFER 1000 m public transport buffer OVERLAY l = pwf * ptb location (using buffer) Figure 1: Cartographic model for ACME site location (final overlay of base map not shown). On the upper left, the model shows how the GIS command RECLASS was used to reclassify the unemployment, education and population data using the classes as defined in section 3.2. The reclassified data are all expressed as factors ranging from 0 to 1. Potential workforce is then calculated using OVERLAY with the mathematical expression pwf = u * e * p. Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 4 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu On the upper right, the model shows how POINTRAS was used to convert the vector data railway station and bus station into raster data. The two raster data sets were then combined into a public transport data set using OVERLAY with the mathematical expression pt = r + b. Subsequently, a buffer zone with a radius of 1000 meters was applied resulting in public transport buffer. Potential workforce was then multiplied with public transport buffer to show the best potential workforce closest to public transportation. Finally, ward boundaries, railways, rail stations and bus stations were added to create the final map. 4 Results 4.1 Intermediate results Figure 2 shows the Potential workforce factor calculated from unemployment, education and population. The green tones represent wards with a high potential workforce. From Figure 2 alone, we would conclude that the two dark green wards, North Reddey and South Reddey, would be most suitable for ACME. Figure 2: Potential workforce factor. Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 5 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu Figure 3 shows the site locations using the potential workforce and the buffer zones from public transport. Again, green tones represent better locations. North Reddey and South Reddey appear to be unsuitable because they lack public transportation. Breddington, with two dark green circles, is now most suited. Figure 3: Overlay of Potential workforce factor and Public transport buffer. Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 6 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu 4.2 Final results The final results of our analysis are presented in Figure 4. Breddington N increasing potential workforce ward boundary railway rail station bus station 5 km Figure 4: Map of Anytown showing Breddington as the best ward for locating ACME's new headquarters. Breddington has both good public transportation and a high potential workforce. Circles denote areas within 1000 meters of either a rail station or a bus station. Green colors indicate areas with a high potential workforce; yellow colors represent areas with a low potential workforce. Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 7 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu 5 Discussion Fuzzy criteria The criteria set by ACME’s management are rather fuzzy. We should have  How is a good communication network defined? Do we need both road and rail interviewed the management asking questions like:  What does good public transport mean? If “good” means “by rail”, than we should links, or is one of the two sufficient?  When is the potential workforce “in close proximity”? If “close” means have left the bus stations out of our analysis. “anywhere within the boundaries of Anytown” than our analysis could have been much simpler. Data limitations There was no suitable ratio dataset to establish true numbers of potential workers living in a ward. Job seekers could have been used instead of Unemployment, if only this data set was more recently acquired than in 1998. Furthermore, level of education, whether ordinal or ratio in nature, is not good enough. We should have had percentages for a range of different education levels, e.g. 5% university, 10% college, etcetera. The lack of recently acquired ratio data on education and unemployment forced us to work with ordinal data. We have reclassified the Education and Unemployment datasets into classes of equal class width and assigned a weight factor to each class. This operation gave us a means to calculate a potential workforce factor. However, the resulting potential workforce dataset is still ordinal. This implies that in a ward with a potential workforce factor of 0.8, the potential workforce is not twice that of a ward with a potential workforce factor of 0.4. Still, the potential workforce dataset clearly shows which wards have a high and which have a low potential workforce. The available datasets have numbers or ranks per ward. This implies that the datasets suffer from both the modifiable areal unit problem and from ecological fallacy. The modifiable areal unit problem means that another administrative subdivision of Anytown would result in a different map and possibly highlights another ward as the best location for ACME’s headquarters. Ecological fallacy results from the assumption that the data are uniformly distributed within the ward. For example, we assume that each smaller portion of a ward (block, street) has the same potential workforce as the whole ward. The limitations mentioned above imply that we can only suggest which ward is best suited for ACME’s headquarters. For a search within a ward, we would need datasets on smaller administrative units like blocks or streets. Furthermore, we lack data on available sites for sale or for rent. Technical problems We have encountered only a few minor technical problems with the datasets and the software package used. These are not worth mentioning here. Recommendations for future research Future research should: Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 8 of 9 ACME International Industries – Site location in Anytown region Jan Stafleu  Start with interviewing ACME’s management to better establish the search  criteria;  Use recently acquired ratio data on unemployment; Use recently acquired ratio data on education, specifying a range of education  levels (university, college, etcetera);  Use recently acquired data on available and suitable sites for sale or for rent; Use datasets on smaller administrative units than wards. Module 2 - ACME International Industries – Site location in Anytown region - Jan Stafleu d20030126.doc Page 9 of 9