Kevin Ohashi
Founder, Review Signal LLC
Freelance Software Developer
Amateur Digital Nomad
Board Member, Affinity Lab
Playing with Data for Fun
and Profit
Who Am I?
Background:
● MSc in Entrepreneurship
● MSc in International Marketing & Brand
Management
● BA in Economics
● Minor in Computer Science
My First Business - Domain Names
Which Domain Names are Valuable?
Source: ICANN.org
Hint: it’s
.COM
What Makes Domain Names Valuable?
● Brandability
● Industry/Market
● Length (shortness)
● Assocations/imagery
● …
● -------------------------->
TRAFFIC!
Domain Name Business Model
Traffic == Money
1. Buy Domain
Names with Traffic
2. Redirect them to
Advertising
3. Collect Checks
Disclaimer: Might be evil. Many
people hate these sites. I don’t blame
them.
$$
$$
Data back in my day...
Yahoo! Overture
● Showed every query’s
popularity
● Updated monthly
● Unlimited Querying
● ==Data Mining Jackpot
What the data taught me...
Data showed a strong correlation between search
volume and traffic
Finding valuable domains was limited by:
● Algorithms to generate names
● Registering them faster than anyone else
Automation is a form of innovation
My Master’s Thesis - Predicting Box
Office Sales with Twitter Data
Why did I do this?
Why Not?
The goal of my research was to show electronic
word of mouth (WoM) == normal word of
mouth.
The byproduct was electronic WoM can be
recorded and analyzed. So I explored the
data...
I Love Movies, So I Studied Them
“The highest adjusted R^2 (.96) for week 3
used: Positive Buzz, Theaters and P/N
Ratio, to explain box office sales. “
tl;dr: The volume of positive
messages, the number of
theaters playing it, and the
ratio of positive to negative
messages explained 96% of
box office sales.
Most Important Data Skill Learned:
Sentiment Analysis
Commercialization of Thesis
How Does Review Signal Work?
Core Business: Sentiment Analysis
● Automatically analyzed over 15 million
opinions
● Created a source of reviews where none
existed before
● Always Up to Date (or Up to Data)
Thank You!
Remember...
● Data can be fun.
● Data can be profitable.
● Data can unlock new information/insights.
● Data is only as good as the person analyzing it.
@KevinOhashi Personal Website:
KevinOhashi.com
Business Website:
ReviewSignal.com
Kevin@ReviewSignal.com

Playing with Data for Fun and Profit

  • 1.
    Kevin Ohashi Founder, ReviewSignal LLC Freelance Software Developer Amateur Digital Nomad Board Member, Affinity Lab
  • 2.
    Playing with Datafor Fun and Profit
  • 3.
    Who Am I? Background: ●MSc in Entrepreneurship ● MSc in International Marketing & Brand Management ● BA in Economics ● Minor in Computer Science
  • 4.
    My First Business- Domain Names
  • 5.
    Which Domain Namesare Valuable? Source: ICANN.org Hint: it’s .COM
  • 6.
    What Makes DomainNames Valuable? ● Brandability ● Industry/Market ● Length (shortness) ● Assocations/imagery ● … ● --------------------------> TRAFFIC!
  • 7.
    Domain Name BusinessModel Traffic == Money 1. Buy Domain Names with Traffic 2. Redirect them to Advertising 3. Collect Checks Disclaimer: Might be evil. Many people hate these sites. I don’t blame them. $$ $$
  • 8.
    Data back inmy day... Yahoo! Overture ● Showed every query’s popularity ● Updated monthly ● Unlimited Querying ● ==Data Mining Jackpot
  • 9.
    What the datataught me... Data showed a strong correlation between search volume and traffic Finding valuable domains was limited by: ● Algorithms to generate names ● Registering them faster than anyone else Automation is a form of innovation
  • 10.
    My Master’s Thesis- Predicting Box Office Sales with Twitter Data
  • 11.
    Why did Ido this? Why Not? The goal of my research was to show electronic word of mouth (WoM) == normal word of mouth. The byproduct was electronic WoM can be recorded and analyzed. So I explored the data...
  • 12.
    I Love Movies,So I Studied Them
  • 13.
    “The highest adjustedR^2 (.96) for week 3 used: Positive Buzz, Theaters and P/N Ratio, to explain box office sales. “ tl;dr: The volume of positive messages, the number of theaters playing it, and the ratio of positive to negative messages explained 96% of box office sales.
  • 15.
    Most Important DataSkill Learned: Sentiment Analysis
  • 16.
  • 17.
    How Does ReviewSignal Work?
  • 18.
    Core Business: SentimentAnalysis ● Automatically analyzed over 15 million opinions ● Created a source of reviews where none existed before ● Always Up to Date (or Up to Data)
  • 19.
    Thank You! Remember... ● Datacan be fun. ● Data can be profitable. ● Data can unlock new information/insights. ● Data is only as good as the person analyzing it. @KevinOhashi Personal Website: KevinOhashi.com Business Website: ReviewSignal.com Kevin@ReviewSignal.com