Data for: Local fuzzy geographically weighted clustering: A new method for geodemographic segmentation.

This dataset (compressed rar file) includes the Matlab code files for "Local Fuzzy Geographically Weighted Clustering" algorithm and a shapefile containing socio-demographic data and cancer incident data across 973 block groups in Manhattan, New York.

The files are:

1. LFGWC.m = The Matlab code of LFGWC (Local Fuzzy Geographically Weighted Clustering)
2. LFGWC_Call.m = The file to run the above code
3. validity.m = The Matlab code for validating the clustering output
4. licence.txt = The file describing the license terms
5. Demo = Dataset folder

Demo folder contains the following:

1. Data.txt = The non-normalized dataset
2. Population.txt = Population for each polygon
3. Distance.txt = Distance among all objects
4. Centroid.txt = Initial cluster centres
5. Shapefile: Manhattan_Data.shp

The shapefile has been originally downloaded from a benchmark dataset of small-area cancer incidence (Boscoe et al. 2016). The benchmark dataset includes 524,503 tumors across 13,823 block groups for the entire New York State diagnosed between 2005 and 2009 (download link: https://www.satscan.org/datasets/nyscancer/index.html).

Manhattan_Data.shp shapefile includes only the county of Manhattan and not the entire NY. Data have undergone slight modifications that are explained in detail in the paper.

Attributes of Manhattan_Data.shp:

DOHREGION Geographic identifier
CODE Unique ID code for joining data
POPULATION Total population (2010 Census)
White_Pop % white alone population (2010 Census)
Black_Pop % black alone population (2010 Census)
Asian_Pop % Asian alone population (2010 Census)
Other_Pop % other race population (2010 Census)
Hispanic % Hispanic population (2010 Census
HH_Size Persons per household (2010 Census)
LT_HS % population less than high school education (25 & over)
Under_Pov % population under poverty (2006-2010 ACS Data)
BC_Rate Incidents of breast cancer per 1000 people
PC_Rate Incidents of prostate cancer per 1000 people
TC_Rate Total cancer incidents per 1000 people

For more information on the original benchmark dataset visit:
https://www.satscan.org/datasets/nyscancer/index.html