'

How Comcast uses Data Science to Improve the Customer Experience

Понравилась презентация – покажи это...





Слайд 0


Слайд 1

• • • •


Слайд 2

• • • • • • •


Слайд 3

• •


Слайд 4

Similarity • 4 3.5 3 2.5 2 1.5 1 0.5 0 0 168 336 504 Δ Hours • • 672 840


Слайд 5

Similarity • 4 3.5 3 2.5 2 1.5 1 0.5 0 0 168 336 504 672 840 Δ Hours Daily pattern: views in similar hours of the days are more similar • •


Слайд 6

Similarity • 4 3.5 3 2.5 2 1.5 1 0.5 0 0 168 336 504 672 840 Δ Hours Weekly pattern: views in similar days of the week are more similar • • Daily pattern: views in similar hours of the days are more similar


Слайд 7

Similarity • 4 3.5 3 2.5 2 1.5 1 0.5 0 0 168 336 504 672 840 Δ Hours Overall decay: recent views are more similar • • Weekly pattern: views in similar days of the week are more similar Daily pattern: views in similar hours of the days are more similar


Слайд 8

0.1 0.06 0.04 0.02 0 0 168 336 ∆ Hours VOD Daypart View Similarity Similarity Similarity 0.08 0 0 168 336 504 Δ Hours 672 840


Слайд 9

Repeated-Search Pattern Repeat percentage 0.4 0.3 0.2 0.1 0 0 168 336 ∆ Hours VOD Daypart View Similarity 0.1 Similarity Similarity Linear(Station-level) View Similarity 0 0 168 ∆ Hours 336 4 3 2 1 0 0 168 336 504 672 840


Слайд 10

• •


Слайд 11

= New Signal 0.7 7 Number of Signals • •


Слайд 12

100% Only Social 80% 60% Nielsen + DVR 40% 20% Nielsen + DVR + Social 0% Top 5 • • Top 10 Top 20


Слайд 13


Слайд 14


Слайд 15


Слайд 16


Слайд 17

Recall vs. Resolution 70.00 65.00 BaseLine 60.00 CoMPASS 55.00 CoMPASS++ Various internal ViaccessOrca DigitalSmith and Rovi external ContentWise Algorithms Recall 50.00 45.00 40.00 35.00 30.00 Gravity 25.00 MortarData 20.00 ThinkAnalytics 0.40 0.60 0.80 1.00 1.20 1.40 1.60 Resolution (Weighted Sum, Normalized by User Study) Compass (Trend)


Слайд 18


Слайд 19


Слайд 20


Слайд 21

• • • •


Слайд 22

• • • •


Слайд 23

• • • •


Слайд 24

• • • •


Слайд 25

Picture courtesy of Cisco


Слайд 26


Слайд 27

• • •


Слайд 28

Network BW Saved by Cache Baseline w/ 100% cache size


Слайд 29

Network BW Saved by Cache Baseline w/ Prediction w/ 100% cache 30% cache size size


Слайд 30

Network BW Saved by Cache Baseline w/ Prediction w/ 100% cache 30% cache size size Prediction w/ 60% cache size


Слайд 31

       


Слайд 32

       


Слайд 33


Слайд 34


×

HTML:





Ссылка: