Nissim Panchpor Arunabh Satpathy Gaurav Gada TECHNOLOGY FEATURES Time scaling Company Comparisons Sentiment analysis Tokenization Heuristics • Ways to identify news article entities. • Count score/total count (normalized) • Exponential decay score = exp(-c/t) • Exponential Decay Score + Count Score + Title Score • A natural language processing, text analysis technique to extract and quantify subjective information. • Valence Aware Dictionary and sEntiment Reasoner (VADER) • Sentence-level aggregation (over article level) • A pipeline to break the data into sentences, then run through the VADER. Deep insights PitchBook integration PROCESS 2.1 million news articles Data cleaning Wireframes User research Lexical sentiment analysis Final product UI/UX & visualization prototypes INVESTING IS EMOTIONAL 1 2 3 4 5 SENTISAGE DETECTS AND DISPLAYS MARKET SENTIMENT THROUGH SENTIMENT ANALYSIS MARKET SENTIMENT IS THE AGGREGATE INVESTOR ATTITUDE TO MARKET CHANGES MARKET SENTIMENT AFFECTS INVESTMENT DECISIONS NEWS ARTICLES ARE USEFUL TO DETECT MARKET SENTIMENT ATALLIC SentiSage Sponsored by

1 2 3 4 5 - University of Washington POSTER.pdfPositive Sentiment +2.5 Negative Sentiment -1.5 SimilarWeb Unique Visitors Social 0.04% Sentiment Web 0.23% Facebook Likes Twitter Followers

Download PDF Report

Upload
others
View
4
Download
0

Embed Size (px)

Citation preview

Page 1: 1 2 3 4 5 - University of Washington POSTER.pdfPositive Sentiment +2.5 Negative Sentiment -1.5 SimilarWeb Unique Visitors Social 0.04% Sentiment Web 0.23% Facebook Likes Twitter Followers

Nissim Panchpor Arunabh Satpathy Gaurav Gada

TECHNOLOGY

FEATURES

Time scaling

Company Comparisons

Sentiment analysis TokenizationHeuristics

• Ways to identify news article entities.

• Count score/total count (normalized)

• Exponential decay score = exp(-c/t)

• Exponential Decay Score + Count Score + Title Score

• A natural language processing, text analysis technique to extract and quantify subjective information.

• Valence Aware Dictionary and sEntiment Reasoner (VADER)

• Sentence-level aggregation (over article level)

• A pipeline to break the data into sentences, then run through the VADER.

Deep insights

PitchBookintegration

PROCESS

2.1 million news articles

Data cleaning

WireframesUser research

Lexical sentiment analysis

Final product

UI/UX & visualization prototypes

INVESTING IS EMOTIONAL

1 2 3 4 5SENTISAGE DETECTS AND DISPLAYS MARKET SENTIMENT THROUGH SENTIMENT ANALYSIS

MARKET SENTIMENT IS THE AGGREGATE INVESTOR ATTITUDE TO MARKET CHANGES

MARKET SENTIMENT AFFECTS INVESTMENT DECISIONS

NEWS ARTICLES ARE USEFUL TO DETECT MARKET SENTIMENT

ATALLIC

SentiSageSponsored by

Similarweb 2016: new tricks and treats by Boaz Sasson

Marketing

Radiation level map of Koriyama City (November, …...Radiation level map of Koriyama City (November, 2013) Reference: August, 2011 less than 0.23 ･･･ 456 0.23-0.29 ･･･

Documents

Modern Keyword Research for Lazy SEOs by Boaz Sasson (Similarweb)

Marketing

Practical Sentiment Analysis Tutorial - BIUu.cs.biu.ac.il/~89-680/sentiment.pdf · Practical Sentiment Analysis Tutorial Jason Baldridge @jasonbaldridge Sentiment Analysis Symposium

Documents

GRU: Taming a Herd of Wild Servers - Oz Katz, Similarweb - DevOpsDays Tel Aviv 2016

Technology

New tricks and treats of Similarweb 2016

Marketing

DLP230GP 0.23 qHD S246 Datasheet datasheet

Documents

$sept 2007 CardMaps · 2017. 7. 20. · you are very pager-naps pagernaps vaeœl\ues sentiment sentiment sentiment sentiment se co hello sent Th ema s sentiment sentiment rñdÞS sentime$