A.Frank Future Gazing Search Engines רואים את הנולד: מנועי חיפוש צופי עתיד Ariel J. Frank Department of Computer Science Bar-Ilan University, Israel [email protected].

A.Frank

Future Gazing Search Engines : צופי חיפוש מנועי הנולד את רואים

עתידAriel J. Frank

Department of Computer Science

Bar-Ilan University, Israel

[email protected]

2 A.Frank

Contents

• Past, Present, …

• Google Suggest/Trends/Insights

• Flight delay and fare casting

• Sentiment Analysis based search

3 A.Frank

Rrrewind

• What if you missed checking what was hot on social media yesterday?

• Rrrewind is a site that let you go back in time and see what was popular.

• It archives top social media topics from numerous social networks: – Delicious, Digg, Hacker News, Reddit, Hulu, Yahoo videos,

YouTube, Dribble, Flickr, Amazon, and Yahoo Buzz.

• Provides original links to contents.• “A Wayback Machine for Social Media“ (Lifehacker)

4 A.Frank

Rrrewind

5 A.Frank

Tweetbeat (Kosmix)

• Lets you know what people are talking about in social networks as it is happening.

• Uses Kosmix’s advanced semantic technology to identify the meaning of social media updates (Twitter, Facebook, YouTube, etc.) in real-time and categorize them instantly.

• This sophisticated categorization system is based on the web’s largest taxonomy that enables Tweetbeat to understand relationships, influence of contributors, and trends so that it finds relevant updates without relying on hashtags/keywords.

• It then uses a relevance meter and a real-time influence score to determine what content to show, and what to filter out.

6 A.Frank

Tweetbeat (Kosmix)

7 A.Frank

Tweetbeat (Kosmix)

8 A.Frank

Future Gazing Search (Engines)?

9 A.Frank

Contents

• Past, Present, …

• Google Suggest/Trends/Insights

• Flight delay and fare casting

• Sentiment Analysis based search

10 A.Frank

Google Suggest

11 A.Frank

Google Suggest

12 A.Frank

Google Instant

13 A.Frank

Google Instant

14 A.Frank

Google Trends

15 A.Frank

Google Trends

16 A.Frank

Google Flu Trends

• Tool released by Google researchers, in 2008, for predicting how many people were getting sick with the flu in different places around the world.

• Google Flu Trends uses aggregated Google search data to estimate current flu activity around the world in near real-time.

• At the time, the tool could predict the likely number of cases in parts of the United States 10 days before the Centers for Disease Control and Prevention (CDC) could.

17 A.Frank

Google Flu Trends

18 A.Frank

Google Flu Trends

United States Flu Activity

Influenza estimate Google Flu Trends estimate  

United States data

United States: Influenza-like illness (ILI) data provided publicly by the U.S. Centers for Disease Control.

19 A.Frank

Google Books Ngram Viewer

• Usage of words and phrases in book publications has developed over the years and it reflects a lot about the progression of societies.

• Provides visual view usage of words and phrases in books over a period of time.

• With Google transforming every book that they can get their hands on into digital format in their Google Books database, it is now a lot easier to conduct these types of search/studies.

• It is a search tool that allows you to see the development of how words have been used in books over the years as displayed in a visual graph.

20 A.Frank

Google Books Ngram Viewer

21 A.Frank

Google Insights for Search

22 A.Frank

Google Insights for Search

23 A.Frank

TrendsBuzz

24 A.Frank

Contents

• Past, Present, …

• Google Suggest/Trends/Insights

• Flight delay and fare casting

• Sentiment Analysis based search

25 A.Frank

Delaycast

• Helps you obtain reasonable estimates of the flight delays you may experience.

• “On-time prediction engine” – estimates are based on factors such as the airline, flight origin and destination, developing trends, holidays, date and time of the flight and so on.

• Flight delay predictions are produced by the models and thus aren’t 100% accurate. Nevertheless, the additional insight provided beyond the airlines schedules should come in handy for most of the travelers.

26 A.Frank

Delaycast

27 A.Frank

Delaycast

28 A.Frank

Farecast (Bing Travel)

29 A.Frank

Farecast (Bing Travel)

30 A.Frank

Farecast (Bing Travel)

31 A.Frank

Farecast (Bing Travel)

32 A.Frank

Contents

• Past, Present, …

• Google Suggest/Trends/Insights

• Flight delay and fare casting

• Sentiment Analysis based search

33 A.Frank

Sentiment Analysis (Opinion Mining)

• Refers to a broad area of data analysis, natural language processing, computational linguistics and text mining.

• Sentiments or emotions get expressed through ideas, opinions, criticisms, and sarcasm. It is measured against what’s positive, what’s negative, and what remains neutral.

• The attitude of an author may be their judgment or evaluation, their affective state (i.e., the emotional state of the author when composing) or the intended emotional communication (i.e., the emotional effect the author wishes to have on the reader).

• Sentiment Analysis is exceptionally challenging as computers are very poor at understanding irony, subtext, implicit context and all kinds of other linguistic constructs that a human easily comprehends.

34 A.Frank

What does the Internet think?

35 A.Frank

The Stock Sonar

• Continuously scans the traditional and social media, bringing up timely insights to the attention so as to be able to see the earliest ripples before the wave hits.

• It provides a one-glance quantified assessment of the financial, news and social media sentiment relevant to stock market analysis.

• Enables rapid, timely and informed trading decisions based on real-time and historical media sentiment on specific companies and sectors.

• The Stock Sonar leverages advanced information extraction technologies developed by its parent company, Digital Trowel.

36 A.Frank

The Stock Sonar

37 A.Frank

The Stock Sonar

38 A.Frank

Recorded Future

• A comprehensive source of past, planned, and predicted future (speculative events) on the web.

• “Temporal and predictive analytics engine“ – extracts time and event information from the web and provides a service for tracking and predicting future trends.

• Statistical analysis of data is used to help predicting future trends to give us insight into what will most probably happen.

• Through analyzing past events and their frequency it can be determined if and when they will reoccur.

39 A.Frank

Recorded Future

40 A.Frank

Recorded Future

41 A.Frank

Recorded Future

42 A.Frank

Recorded Future

43 A.Frank

Recorded Future

44 A.Frank

Recorded Future

http://blog.recordedfuture.com/tag/predictions/

45 A.Frank

Recorded Future

Bartz and Yahoo - 6 Month Negative Sentiment

http://blog.recordedfuture.com/tag/predictions/

46 A.Frank

Recorded Future

Bartz and Yahoo - 6 Month Positive Sentiment

http://blog.recordedfuture.com/tag/predictions/

47 A.Frank

Recorded Future

48 A.Frank

Recorded Future

49 A.Frank

Recorded Future

50 A.Frank

Recorded Future

51 A.Frank

Recorded Future

52 A.Frank

Still time for future SEs

53 A.Frank

Still time for questions?