Geodemographics and Big Data

Selected slides only

Transcript

1. Guy Lansley Department of Geography, UCL g.lansley@ucl.ac.uk @GuyLansley Geodemographics and Big Data: An Introduction to Geo-temporal Demographics Map: Oliver O’Brien
2. What this talk will cover • Introduction to my research • Population data • Geodemographics & population geography • Geo-temporal demographics
3. Guy Lansley Education • 2008-11 BA Geography – University of Sheffield • 2011-12 MSc GIS – University of Leeds • 2014+ PhD (part-time) – UCL Positions • 2012-14 Teaching Fellow Department of Geography, UCL • 2014+ Research Associate Department of Geography, UCL Consumer Data Research Centre
4. My research interests Quantitative Human Geography Geodemographics Big Data Geo-temporal Demographics Social Media Data Retail Geography Consumer Data Visualisation GeographicInformation Science (GIS) Text mining Social inequalities
5. GEODEMOGRAPHICS Statistical data relating to the population and particular groups within it. Geographic
6. BIG DATA Large and complex data sets
7. The 3 V’s of Big Data Volume Variety Velocity
8. Geographic data and us • Geographic information is becoming more integral with daily life – Services, websites & apps
9. Population Data
10. Who needs small area population statistics?
11. Sources of UK population data Censuses and official statistics Administrative records Transport data Retail data Social media Mobile phones
12. Geodemographics
13. People are not randomly distributed within cities
14. Inequality: House prices
15. Transport • Understanding where the population travel to work Visualisation of where people travel to work (data source: 2011 Census)
16. Why do people cluster?
17. Models of Urban Morphology Concentric rings model of Chicago’s neighbourhoods (1925)
18. Geodemographics today • The 2011 Output Area Classification (OAC) • Produced from 2011 Census • 60 variables • Produced 8 Supergroups public.cdrc.ac.uk
19. Instead of just social class Geodemographic classifications group units* by a range of population and local characteristics * The units could be neighbourhoods (such as Output Areas) or even individual households if data is available Travel
20. Geo-temporal Demographics
21. Lots of data on where we live What do you do at home daily? 8 Hours sleeping? 4 hours watching television? (Ofcom) Most geodemographic research is focused on where people live. However, 100% of the populations time is not spent at their recorded homes Home
22. Geo-temporal Space Time
23. Twitter A small proportion of Twitter users make their location publically available
24. Forenames - Gender
25. Forenames – Age (Females) 5 clusters of forenames based on their age distributions
26. What about the content of Tweets?
27. What are people saying? • Tweets are influenced by space: – Place – Land use – Activity – Sentiment – Language Note: the words “Greater” and “London” have been removed Soho • Tweet messages are also influenced by time and date
28. Topic Modelling • A means of understanding general trends in text data, so we can analyse how the content of Tweets may vary by space and time • To test this I generated 20 topics from 1.5 million Tweets in Central London
29. Guy Lansley Department of Geography, UCL g.lansley@ucl.ac.uk @GuyLansley End