Towards a Comparative Science of Cities: Using Mobile Traffic Records in New York, London, and Hong Kong - Computational Approaches for Urban Environments - page 367

a b

Core Business
Commercial
Mixed
Residential
Airports
Parks

HD Res.
LD Res.
Business
Mixed
Infra
Parks
Other

c d

2.5 1.5

SMS SMS

0. 1

2

1.5

0

1

−0.5

0.5

−1

0

3 1

Calls Calls 2.5

0.5

2

1.5 0

1

−0.5

0.5

0 −1

2.5 1

Request Request

2

0.5

1.5

0

1

−0.5

0.5

0 −1

2.5 1

UL data UL data 2

0.5

1.5

0

1

−0.5

0.5

0 −1

2.5 1

DL data DL data

2

0.5

1.5

0

1

−0.5

0.5

0 −1

Mon Tue Wen Thu Fri Sat Sun Mon Tue Wen Thu Fri Sat Sun

Fig. 15.7 New York clusters .( a ) Spatial projection of K D 6 clusters, with their interpretation
in the legend (see details in main text). ( b ) Actual land use maps as extracted from census data. ( c )
Signatures of the clusters in the different components of activity. ( d ) Deviations of the signatures
compared to the whole city signatures displayed on Fig. 15.4 . Colors on the signature plots match
those on the cluster map ( a ), and gray areas correspond to zones with no recorded data

population and job densities within the clusters (when available), as well as
confusion matrices highlighting the similarities between our clusters and the land
use classes.
A few patterns are easy to interpret and compare with census data. For exam-
ple:

￿ Cluster 1 (in red). In all cities, these clusters' signatures present high levels of

activities during working hours and very low levels of activity in the evening