Putting Manual Cartographic Techniques Back into the Digital Era - A Python-based Algorithm for Enhanced Dot Density Mapping
- 1. A Python-based Algorithm for Enhanced Dot Density Thematic Mapping Daryn Hardwick Saint Cloud State University Department of Geography & Planning
- 2. The Problem Expectations/Hypothesis Background • Python • Center of Gravity Principle • Improvements to the Dot Density Technique • Anonymity in Agricultural Data Methods • Algorithm • Test on Agricultural Data Results Additional Functionality Discussion/Conclusion • Critiques of the Algorithm • Future Research
- 3. Randomness in Dot Density Thematic Mapping • Use of GIS Software Tobler’s First Law of Geography • “…near things are more related than distant things” Previous Solutions • Drawing Programs • Feature masks
- 4. Expectation • The algorithm presents a solution to the problem of randomness using the center of gravity principle Hypothesis • The dots produced from the algorithm will be significantly closer to actual farmland than the random dot distribution produced by GIS software
- 5. What is Python? • An object-orientated scripting language How does it work? • Methods (procedures) performed on or between objects1 • In GIS, objects are data with properties Why Python? • ArcGIS compatibility • Software quality, developer productivity, program portability, support libraries, component integration, and enjoyment2 1: (Stefik and Bobrow 1985) 2: (Lutz 2008)
- 6. What is the center of gravity principle? • “…the cartographic ideal, back when all dots were placed manually, has been to locate the dots as close to the real distribution as possible.”1 Algorithm provides solution • Placement of known points 1: (Dent et al. 2009)
- 7. Earliest dot maps 1 • 1852 – cholera maps by August Petermann • 1863 – Maori population in New Zealand Percentage Dot Maps 2 Limiting amount of dot overlap 3 The “Fuzzy Dot Map” 4 1: (MacEachren 1979) 2: (Mackay 1953) 3: (Kimerling 2009) 4: (Alqvist 2009)
- 8. USC (Title 7, Chapter 55, § 2204g) 1 • “…information obtained may not be used for any purpose other than statistical purposes for which the information is supplied.” Algorithm does preserve anonymity 1: (Department of Agriculture 2008)
- 9. Critical Inputs • Enumeration unit areas • Known point locations • Two Geodatabases • Areas to be masked from receiving dots (i.e. Water) • Dot value Additional Inputs • Clustering • Number of Buffers
- 10. Split of enumeration areas Area and buffer distances are calculated
- 11. Why use variable width buffers?
- 12. Split of enumeration areas Area and buffer distances are calculated Buffers clipped Number of output dots calculated Buffered areas merged Enumeration areas merged Areas excluded from output dots masked Creation of the output dot map
- 13. 2007 Census of Agriculture • Acreage, number of farms, market value of ag. products sold Dots created using the algorithm and random dot placement tested Near analysis Significance Testing Land Cover data retrieved from the 2006 National Land Cover Dataset
- 14. Market Value Acreage of Number of Of Agricultural Farmland Farms Products Sold
- 15. Random Algorithm One dot represents 12,800 acres 68.33% On Farmland 70.14% 141.37 meters Distance to 103.39 meters Farmland
- 16. Random Algorithm One dot represents 40 farms 63.01% On Farmland 65.31% 189.01 meters Distance to 125.82 meters Farmland
- 17. Random Algorithm One dot represents $5,000,000 72.51% On Farmland 76.39% 66.38 meters Distance to 38.82 meters Farmland
- 18. T-score P-value Significant? Acreage: -1.93 0.0269 Yes Number of farms: -2.72 0.0033 Yes Market value: -3.74 0.0001 Yes
- 19. Operation within ArcGIS Custom Toolbox Two Parts • Step 1 – Script • Step 2 - Model
- 20. Algorithm run with a low clustering effect Algorithm run with a high clustering effect
- 21. Algorithm run with a two buffers Algorithm run with a four buffers
- 22. Critiques • Time • Placement of Known Points Future Research • Improve the issue of time • Another way to solve this problem?
- 23. This algorithm • Increases the accuracy of placed dots • Adheres to the center of gravity principle • Removes some randomness without sacrificing anonymity of underlying data • Can be used in ArcGIS • Options to further customize the output
- 24. Ahlqvist O., (2009) “Visualization of Vague Category Counts – Introducing the Fuzzy Dot Density Map”. In: International Cartographic Conference 2009 Proceedings, International Cartographic Association. Dent B., Torguson J., and Hodler T., (2009) “The Dot Density Map”. Cartography: Thematic Map Design, 6e: 119-130. Department of Agriculture, (2008) “Authority of Secretary of Agriculture to conduct census of agriculture”. U.S. Code Title 7, Chapter 55, § 2204g. Golledge R., (2002) “The Nature of Geographic Knowledge”. Annals of the Association of American Geographers, 92(1): 1-14. Hey A., (2012) “Automated Dot Mapping: How to Dot the Dot Map”. Cartography and Geographic Information Science, 39(1): 17-29. Hoonaard W., (2003) “Is Anonymity an Artifact in Ethnographic Research?”. Journal of Academic Ethics, 1: 141-151. Kimerling A., (2009) “Dotting the Dot Map, Revisited”. Cartography and Geographic Information Science, 36(2): 165-182. Lutz M., (2008). “A Python Q&A Session”. Learning Python, 3e: 3-20. MacEachren A., (1979) “The Evolution of Thematic Cartography / A Research Methodology and Historical Review”. The Canadian Geographer, 16(1): 17-33. Mackay J., (1953) “Percentage Dot Maps”. Economic Geography, 29(3): 263-266. Stefik M., and Bobrow D., (1985). “Object-Orientated Programming: Themes and Variations”. AI Magazine, 6(4): 40-62. Tobler W., (1970) “A computer movie simulating urban growth in the Detroit region”. Economic Geography, 46(2): 234-340.
- 25. Graduate Student Saint Cloud State University daryn.hardwick@yahoo.com