New Insights into Individual Activity Spaces using Crowd-Sourced Big Data

Presentation to Big Data Science

Stanford, CA, USA, May 27 - May 31, 2014

Dr Nick Malleson and Prof. Mark Birkin
School of Geography, University of Leeds


Context: Towards a model of daily urban dynamics

Background and theoretical frameworks

Shoppers in Hong Kong

Space-time geographies

Awareness / activity spaces

Relevant research

Data and the study area


Estimating anchor points

Textual analysis

Results and Conclusions


Excellent residential (night-time) population data (from censuses etc.).

But limited ambient (day-time) population data.

Ambient population is relevant to diverse fields:

Disease spread / mitigation

Transport planning

Pollution / disease burdens

Crime (e.g. [13])

. . .

Clustering of crime

Research Aims

Evaluate the extent to which crowd-sourced data can be used to gain insight into the individual activity spaces (the places that individuals visit on a regular basis)

Identify the functions of different anchor points

Background: Space-Time Geography

Cities characterised as complex systems (e.g. [1])

Aggregate smoothing of underlying dynamism

Space-time prism

Theories exploring spatio-temporal movement constraints

E.g. Time Geography [2]

Better understanding of:

Who visits particular locations

Which other locations have those people visited

Who might they have met during their journey

Background: Activity Spaces and Anchor Points

First step towards a more realistic model of individual behaviour

Activity spaces

The spaces in which an individual's normal activity occurs

Anchor points

Important individual locations (home, work, school, etc.).

Central pillar for diverse disciplines (e.g. criminology, geography)

Very difficult to quantify

New opportunities offered by 'big' crowd-sourced

Big Data: from a geographer's perspective

Traditional data

UK census: ~2000 output areas in Leeds

Microsimulation: ~800,000 individuals

Our twitter data: ~2M points

Big Data = "It won't open in Excel!"

New data...

Mobile telephones: ~5Bn events daily

Public transport journeys (Oyster): 3.5M daily tube journeys (not including busses, cycle hire, etc.)

As well as volume, there are difficulties with velocity and variety

Big Data: from a geographer's perspective

Similar to fourth paradigm data intensive research [3] in the physical sciences

"Crisis" in "empirical sociology" [4]

"One of the areas that is being most dramatically shaken up by N = all is the social sciences. They have lost their monopoly of making sense of empirical social data, as big data analysis replaces the highly skilled survey specialists of the past. .. When data is collective passively while people do what they normally do anyway, the old biases associated with sampling and questionnaires disappear." [5]

Data and the Study Area

Tweet density

Leeds, UK

Large city in north England

Population 757,655 (in 2012)

Central shopping / leisure / commercial area

Social media data

Geo-located Twitter messages from 22 June 2011 to 14 April 2013

N=1,955,655 (after cleaning)

Total message density for all prolific users

Stage 1 - Extract prolific users

Those with 50+ messages in the data

Example 3D kernel

Stage 2 - Generate a message density surface

Kernel density estimation

Highlight areas with high spatial message density for individual users

Generating message density (3D)

Stage 3 - Identify areas of 'unusually' high density (anchor points)

GIS method used to identify peaks in digital elevation data

Use Landserf free software (Java) [11]

Example of GIS analysis to identify peaks

Anchor points are centres of peaks

Assume highest peak is 'home'

Activity space identification - good results
Activity space identification - bad results

Stage 4 - Analysing place functions

Identify words with distinction spatio-temporal profiles [12]

Calculate 'domestic skew' - proportion of words that occur at 'home' locations compared to anywhere else ('away').


Domestic skew for some marker words

Summary / Conclusions

Explore awareness spaces / anchor points from social media data

Estimate 'home' locations

Some words clearly associated with 'home', others with 'away'

Severe limitations to the data, but future potential is promising.

Best available resource?

Clustering of crime

Future work

Further anlysis. Particularly:

Validation / verification to handle skew, representation, accuracy, bias

Comparison with other data (e.g. geodemogrpahics)

Identify additional activities (shopping, leisure, school, etc.).

National language processing

Leeds Institute for Data Analytics (LIDA) and the Consumer Data Research Centre (CDRC)

Similar Research

Classifying neighbourhoods based on social media contributions

LiveHoods ( [6]

Functional profiles or areas [7]

Land use classification [8]

Identification of user 'patches' (regular activity and function) [9]

Anchor points in mobile phone data [10]


[1] M. Batty, "Agents, cells, and cities: new representational models for simulating multiscale urban dynamics," Environment and Planning A, vol. 37, pp. 1373-1394, 2005.

[2] T. Häagerstrand, "What about people in regional science?" Papers of the Regional Science Association, vol. 24, no. 1, pp. 6-21, 1970.

[3] Bell, G., Hey, T., Szalay, A., 2009. Beyond the data deluge. Science 323, 1297-1298. doi: 10.1126/science.1170411

[4] Savage, M, and R Burrows. 2007. "The Coming Crisis of Empirical Sociology." Sociology 41 (5): 885-899.

[5] Mayer-Schonberger, Viktor, and Kenneth Cukier. 2013. Big Data: A Revolution That Will Transform How We Live, Work and Think. John Murray.

[6] J. Cranshaw, R. Schwartz, J. Hong, and N. Sadeh, "The livehoods project: Utilizing social media to understand the dynamics of a city," in Sixth International AAAI Conference on Weblogs and Social Media 2012-05-20.

[7] F. Kling and A. Pozdnoukhov, "When a city tells a story: urban topic analysis," in Proceedings of the 20th International Conference on Advances in Geographic Infomation Systems, ser. SIGSPATIAL '12. ACM, 2012, pp. 482-485.

[8] V. Frias-Martinez, V. Soto, H. Hohwald, and E. Frias-Martinez, "Sensing urban land use with twitter activity," 2013. [ Online].

[9] Y. Qu and J. Zhang, "Regularly visited patches in human mobility," in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, ser. CHI '13. ACM, 2013, pp. 395-398.

[10] S. Isaacman, R. Becker, R. Caceres, S. Kobourov, M. Martonosi, J. Rowland, and A. Varshavsky, "Identifying important places in people's lives from cellular network data," in Pervasive Computing, ser. Lecture Notes in Computer Science, K. Lyons, J. High- tower, and E. M. Huang, Eds. Springer Berlin Heidelberg, 2011, no. 6696, pp. 133-151.

[11] Wood, J. "Identifying Mountains with GIS". In Heywood, I., Cornelius, S. and Carver, S. An Introduction to Geographical Information Systems. 2011. Prentice Hall

[12] M. Birkin, K. Harland, and N. Malleson, "The classification of space-time behaviour patterns in a british city from crowd-sourced data," in Computational Science and Its Applications - ICCSA 2013 , ser. Lecture Notes in Computer Science, B. Murgante, S. Misra, M. Carlini, C. M. Torre, H.-Q. Nguyen, D. Taniar, B. O. Apduhan, and O. Gervasi, Eds. Springer Berlin Heidelberg, 2013, no. 7974, pp. 179-192.

[13] Malleson, N and M. Andresen. "The impact of using social media data in crime rate calculations: shifting hot spots and changing spatial patterns.Cartography and Geographical Information Society (CaGIS).

Thank you

Dr Nick Malleson, Prof. Mark Birkin
School of Geography, University of Leeds