Mapping data Representing data with maps
Geographic analysis tasks • Mapping where things are • Mapping the most and the least • Mapping density • Finding what’s inside • Finding what’s nearby • Mapping changes in time
Why mapping your data – Maps are more effective than tables or lists to communicate some type of informations – Maps take advantage of our ability to distinguish colors, shapes, and patterns to interpret spatial relationships.
Number of cholera cases during weeks 47-2001 and 9-2002 in Katanga, RDC Kitenge 165 Mulongo 197 Manono 19 Kasenga 26 Mufuga 118 Kabalo 493 Manika 162 Kansenya 242 Ankoro 1381 Kabondo 630 Kikula 679 Bukama 1378 Kambove 38 Malemba 1031 Rwashi 115 Katuba 207 Kikondja 3607 Kenya K 673 Kapolowe 887 Lubumbashi 100 Kampemba 403 Dilala 285 Kamina 4
Number of cholera cases during weeks 47-2001 and 9-2002 in Katanga, RDC
Maps can be • Not geocoded (georeferenced) – Representation of the space without any link with the reality (changes every representation) • Geocoded (georeferenced) – Representation of the reality in a geographical model where objects are link with geographical coordinates (Lat. Long. or address)
Distribution of asthma cases in Barcelona Not georeferenced
Georeferenced map
Geocoding • The process of detemining the coordinates of a specific location based on its street address or its existence within a known region. • Coordinates can be assigned as a pair of XY coordinates or as Lat. and Long.
Georeferenced map ? Yes, because linked with their address
Kinds of maps • Reference maps – Simply convey information about location of geographic features (rivers, streets, houses, HC, Hospitals, etc.) • Thematic maps – detect, describe, and analyse spatial patterns; describe associations and correlations with other variables
Reference map
Thematic maps use patterns, colours, or differently sized symbols to display variations in data.
Thematic maps use patterns, colours, or differently sized symbols to display variations in data.
How to present data in a map • Single Symbol • Graduated Colour • Graduated Symbol • Unique Value • Dot • Chart
How to select the appropriate legend • Magnitude, integers (number • Graduate symbol, of cases per area) Dot density • Graduate color areas • Rates, range of values • Every feature of same origin • Syngle symbol same color (streets, Hospitals,HC, rivers, etc) • Categorical data (countries, • Unique value regions, districts, types of roads • Values of many data • Chart legends attributes at once
Mapping data • The type and format of data and study area to be mapped affect the method of mapping
The area to be mapped
The area to be mapped
The area to be mapped
The area to be mapped The importance of Geographical unit of measurement
The area to be mapped
The area to be mapped depends from: • The choice of boundaries. • The geographical unit • The scale chosen to represent • The subject or data of the study: • Residence • Onset • Exposure • Notification • The disease to be mapped
Area representation problem You need to aggregate data because specific problems; e.g. Confidentiality
Area representation problem Some representation, count, can lose resolution
Area representation problem Different aggregation Can bring to different results
Place of residence
An outbreak of Trichinosis in Paris Cases represented by place of residence
Mapping data requires careful interpretation • Underlying population structure – Population (arbitrary denominator) – Population density(higher density higher expected cases) – Age, sex structure (standardization)
Cumulative cases of cholera From week 39-2001 To week 17-2002 in Katanga Level of representation
Distribution of cases of PERTUSSIS Lebanon, as of week 2003-15 # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # Cases/100,000/Y 0.178 – 0.554 0.555 – 0.872 # 0.873 – 1.741 1.742 – 3.554 No report
Population density inhab/SqKm by district in Katanga
Incidence rate of cholera cases per thousand From week 39-2001 to week 17-2002 in Katanga Selection of denominator
Selection of geographical unit- changing in denominator
Selection of geographical unit- changing in denominator
Finally, what data do you want map • Exact location, individual cases • Aggregated data
Specific cases Geocoded to identify their exact location according to hospital Location in Brooklyn, NY Hospitals Cases
Aggregated meningitis cases in Burkina, 2002, expressed as AR/100000.
Hospital Mapping clusters Case 4 Guangzhou Institute TCM Respiratory Tong Sheng Ji Diseases Clinic De Li Restaurant Restaurant Tea Stall Shared Passageway Case 3 Case 2 (Dec 23) Pearl River
Cases representation Distribution of residents by clinical status Nursing home, Delaware, USA, 1992.
Steps in Surveillance Analysis Place
Descriptive Analysis Place • Dot maps for count of cases • Administrative area maps for rates – Choice of administrative areas – Rates to account for population – Standardised rates to account for population structure • “Raster” maps for sentinel surveillance • GIS when case coordinates available
Notification of Tuberculosis in France, 1996 4-Week Period Ending 31/12/1996
Descriptive Analysis of Place Use of Rates • Count of cases does not represent risk • Administrative areas have different populations • Population may vary over time – Seasons – Population influx (refugees) • Rates allow to compare risk • Choice of administrative areas (Problem of small numbers of cases) • Choice of ranges
Notification Rate of Tuberculosis in France, 1996 Cases/100,000
Descriptive Analysis of Place • Mortality/1000 – Region A, 12.97 – Region B, 7.15 • People dye more in Region A? • Population age structure – Region A, 24.7% over 64 years – Region B, 9.2% over 64 years
Distribution of Death by Falls by Province, Canada, 1998 Crude deaths rate per 100,000 Age Standardized Rate per 100,000
Maps for Sentinel Systems Incidence of diarrhea in France, 1995 Cases / 100,000 population Source: Réseau National Télématique des Maladies Transmissibles
Place Testing for Hypothesis • Remove effect of age – Standardisation if needed • Detect clusters – Different techniques • Eg. Test for spatial correlation by nearest neighbour • Eg. Test for spatial correlation by contiguity analysis • Generate hypothesis about risk factors – Overlaying exposure and outcome – Test for cross-correlation
Mapping diseases • Diseases mapping plays an important role in monitoring community’s health • It is a very useful tool in surveillance and alert • Mapping has at least four major roles for PH – Monitoring the spread of infectious disease in order to identify the cause – Monitoring health service access and use – Generating hypothesis about diseases causation or for identifying clusters – Can incorporate ecological analysis
Power of maps • Using maps to present geographic data allow us to see: – Location and extent – Spatial distribution: Pattern, density and Concentration – Spatial interaction:connectivity, accessibility, agglomeration
Power of maps • Easy to answer these questions – What is the size of the place interested – What is there at this point – What kind of distribution does it make – Is it found throught the study area?
Limitation of maps • Map can only answer WHAT and WHERE but not effectively to answer HOW and WHY • Map is not efficient to answer these questions: – Why has it spread or diffused in this particular way – What geographic factors have constrained its spread – How far would it expand – What else is there spatially associeted with that phenomenon
But…you can empower them: • Maps can effectively empowered if they can: – Be used interactively and turned on or off – Integrate with different types of data (aerial photos, satellite imegeries, attribute table, etc.) – Be superimposed to see the spatial relationships with each other – Create spatial patterns based on surveyed data – Display data according to selection criteria
Recommend
More recommend
