Not All Apps Are Created Equal: Analysis of Spatiotemporal - - PowerPoint PPT Presentation

β–Ά
not all apps are created equal
SMART_READER_LITE
LIVE PREVIEW

Not All Apps Are Created Equal: Analysis of Spatiotemporal - - PowerPoint PPT Presentation

Jan 24, 2023 239 likes β€’412 views

Not All Apps Are Created Equal: Analysis of Spatiotemporal Heterogeneity in Nationwide Mobile Service Usage Cristina Marquez and Marco Gramaglia (Universidad Carlos III de Madrid); Marco Fiore (CNR-IEIIT) ; Albert Banchs (Universidad Carlos III de

slide-1
SLIDE 1

Cristina Marquez and Marco Gramaglia (Universidad Carlos III de Madrid); Marco Fiore (CNR-IEIIT); Albert Banchs (Universidad Carlos III de Madrid and Institute IMDEA Networks); Cezary Ziemlicki and Zbigniew Smoreda (Orange Labs)

Analysis of Spatiotemporal Heterogeneity in Nationwide Mobile Service Usage

Not All Apps Are Created Equal:

1

slide-2
SLIDE 2
  • Current status of mobile services:

– Superficial comprehension – Restricted to a small set of coarse-grained datasets

  • Aim: characterize the usage of

mobile services at a national scale given a large dataset

  • Analysis of traffic behavior of

services

2

INTRODUCTION

Across time & space

Properly dimension &

  • rchestrate the mobile

network Supporting data mining techniques Understanding social behaviors

slide-3
SLIDE 3

AIM: mobile service overview

3

DATASET

Dataset collected at Orange core network 1 week from September 24, 2016 User population ~30 million individuals Distributed over > 550,000 π’π’πŸ‘ Granularity of 5 mins ~25,000 base stations (distributed over > 36,000 communes) ( ~ 16 𝑙𝑛2 each) We aggregated data per commune Data recorded at passive probes at Gn and s5/s8 interfaces of GGSN & P-GW

slide-4
SLIDE 4

4

YouTube: YouTube WEB, YouTube Streaming HTTP, YouTube TLS, YouTube Streaming MP4,YouTube Apple

time commune service ul dl 1475067900 01001 5 3780 151200 1475078100 . 6 26875 328412 1475094840 . 7 21768 715481 1475051700 . 8 5654 111236 1475063520 97424 9 2584 20596

DATASET: DEEP VIEW

service Description 1 YouTube WEB . Instagram . Web Advertising . Wikipedia

500 Shazam

7 macro-category

500 distinct services Selection of 20 main categories (most representative) High granularity! Extensive dataset!

slide-5
SLIDE 5

ANALYSIS

5

slide-6
SLIDE 6
  • Focus on weekly demand for each traffic over communes:

6

TIME SERIES ANALYSIS

Apple Store YouTube Facebook SnapChat

Each time series is characterized by a variety of fluctuation In all cases higher diurnal activity (activity reduced at night). Distinctive dichotomy between weekends & weekdays Different temporal patterns between categories & similar services

slide-7
SLIDE 7

7

ARE THEY REALLY SIMILAR?

K-Shape Time Clustering: check goodness of fit with distinct quality indices vs the #clusters K

  • Davies-Bouldin (top graphs)
  • Dunn, Silhoutte (bottom graphs)

Downlink Uplink All possible k considered! To be minimized To be maximized NOT QUITE SIMILAR! Best option? 19 clusters

slide-8
SLIDE 8

8

PEAKS DETECTED

AppleStore

Same macro- category, different behavior

slide-9
SLIDE 9
  • Significant peaks of activity also in space:

9

SERVICE USAGE GEOGRAPHY

Twitter NetFlix

Similar geographical pattern Except 2 outliers

It is used

  • utdoors

It is ubiquitous

Bytes/ subscriber

slide-10
SLIDE 10

10

INSEE urbanization distribution

DOES THE SPACE HAVE AN INFLUENCE IN TIME DYNAMICS?

slide-11
SLIDE 11
  • Correlation of mobile services for different urbanization levels
  • Each bar shows the average 𝑠2 value.
  • In all cases but TGV, the correlation is extremely high

urbanization level has little impact on temporal dynamics of category usage.

  • Service usage changes when people are aboard TGV.

11

ARE TIME SERIES RELATED?

Depends on the train’s schedule

slide-12
SLIDE 12
  • Slope of least square regression of per-subscriber time series
  • Findings:

– Semi-urban & urban areas present similar individual service usage level – Subscribers in rural areas consume around Β½ of the mobile service data in urban areas – Users on TGV generate on average twice or more volume of traffic than urban areas

12

SIMILAR USAGE IN TERMS BYTES/SUBSCRIBER?

slide-13
SLIDE 13

temporal, spatial & hybrid dynamics of mobile services categorized at a national scale finding new interesting macroscopic properties of traffic

  • Findings:

– No 2 services exhibit similar time patterns – Mobile services have very comparable geographical distributions – The urbanization level influences how users consume mobile services, but limited on when they do so – Unique time dynamics on high-speed trains

13

CONCLUSIONS

We studied

granularity

slide-14
SLIDE 14

Cristina MΓ‘rquez /Dec 13th, 2017/ Not All Apps Are Created Equal

14

Cristina Marquez mcmarque@pa.uc3m.es

slide-15
SLIDE 15

15

slide-16
SLIDE 16
  • Data recorded at passive probes at the Gn and s5/s8 interfaces of GGSN & P-GW
  • DPI techniques classify 88% of the mobile traffic
  • Geo-referencing of the IP sessions by examining ULI (User Location Information)

16

3G/4G NETWORK